Qaman
Qaman
Qaman
1
Quality
Analyst
User’s Manual
(7316)
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
Contents at a Glance
CHAPTER 1
Introduction ..........................................................................................1
Overview of NWA Quality Analyst Features ........................................ 1
Background Knowledge Required ....................................................... 3
Typographical and Layout Conventions .............................................. 3
Technical Support................................................................................ 4
CHAPTER 2
Quality Analyst Files..............................................................................7
File Names ........................................................................................... 7
File Types and Their Extensions .......................................................... 7
Data File Structure ............................................................................ 15
Subgroup Types................................................................................. 17
Missing Data Handling....................................................................... 24
Tagged Data ...................................................................................... 26
CHAPTER 3
Creating and Editing Data Files...........................................................27
Opening an Existing Data File ........................................................... 28
Creating a New Data Set ................................................................... 29
Modifying the Data Structure and Display ........................................ 70
Editing Functions............................................................................... 73
The Data Menu .................................................................................. 90
Inserting Data from Another Data Set ............................................ 112
Saving a Data File............................................................................ 112
Printing Data ................................................................................... 113
CHAPTER 4
Selecting Variables........................................................................... 121
Selecting a Single Variable .............................................................. 121
Selecting Multiple Variables ............................................................ 123
Changing the Display Order ............................................................ 125
CHAPTER 5
SQC Charting and Analysis ............................................................... 126
Selecting a Variable First................................................................. 126
Selecting a Chart Type First ............................................................ 127
Variable Control Charts ................................................................... 128
Process Capability Charting and Analysis ....................................... 137
Attribute Control Charts .................................................................. 141
Pareto Diagrams.............................................................................. 144
CHAPTER 6
Managing Charts with the Graphics Viewer .................................... 187
Running the Graphics Viewer Alone................................................ 188
Configuring the Graphics Viewer Interface..................................... 191
Individual Chart Windows............................................................... 192
Printing Charts and Text Reports .................................................... 194
Annotating Charts ........................................................................... 203
Viewing Chart Details ...................................................................... 206
Assigning Causes and Corrective Actions to Data Points ................ 217
Configuring the Appearance of Graphics and Text.......................... 220
Saving Charts and Reports .............................................................. 230
Dragging and Dropping ................................................................... 232
Copying Charts and Text Reports to Other Programs ..................... 233
CHAPTER 7
Customizing Charts .......................................................................... 236
System Parameters ......................................................................... 242
File Parameters ............................................................................... 278
Variable Parameters........................................................................ 283
Control Chart Parameters................................................................ 295
Process Capability (Histogram) Parameters ................................... 318
Capability Report Parameters ......................................................... 331
Box Plot Parameters........................................................................ 332
Pareto Parameters .......................................................................... 333
Short Run Parameters ..................................................................... 336
CHAPTER 8
General Statistical Analysis.............................................................. 338
Overview ......................................................................................... 338
One-way ANOVA.............................................................................. 340
Chi-square ....................................................................................... 344
Correlation....................................................................................... 349
Kruskal-Wallis Statistic ................................................................... 351
Descriptive Statistics....................................................................... 353
t-Statistics ....................................................................................... 356
Wald-Wolfowitz Runs Test .............................................................. 364
Regression Analysis......................................................................... 366
iv • Contents at a Glance
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
CHAPTER 9
Automating Procedures with Run Files............................................ 384
Creating Run Files ........................................................................... 385
How Run Files Work ........................................................................ 385
Run File Commands and Parameters .............................................. 386
Replaceable Parameters.................................................................. 484
Creating Complete Task Sequences ................................................ 490
Creating Header Files from within Run Files ................................... 495
Executing Run Files ......................................................................... 496
Run File Errors................................................................................. 505
Using Older Run Files ...................................................................... 506
CHAPTER 10
Using External Data Files ................................................................. 510
Importing Data................................................................................ 510
Exporting Data to a Comma-Delimited File..................................... 521
Using the Clipboard with the Editor ................................................ 522
CHAPTER 11
Manipulating Data Files ................................................................... 530
Changing the Missing Data Symbol................................................. 530
Rotating a Data Set ......................................................................... 531
Converting Subgroups To or From a Single Column........................ 532
Extracting Subsets from a Data File ................................................ 536
Defining Calculations for Run files .................................................. 538
CHAPTER 12
Database Connectivity ..................................................................... 542
Overview ......................................................................................... 542
Before Using Connectivity ............................................................... 542
Defining Connectivity Data Sets...................................................... 543
Using Replaceable Parameters with Filter Criteria in SQL
Statements ...................................................................................... 573
Using Connectivity Data Sets .......................................................... 579
Modifying Connectivity Data Sets ................................................... 580
Using Database Connectivity in Run files........................................ 583
Configuring and Troubleshooting.................................................... 587
Getting Technical Support for Connectivity .................................... 599
APPENDIX A
Missing-Data Techniques in Control Charts..................................... 600
Missing Data with Rowwise Subgroups .......................................... 601
Missing Data with Columnwise Subgroups ..................................... 606
Contents at a Glance • v
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
APPENDIX B
Alternative Statistical Calculation Methods..................................... 624
EPA Control Limits........................................................................... 624
Ford Process Capability Indices ...................................................... 625
Control Chart Limits Based on Probability ...................................... 626
Standard Deviation Calculation Methods ........................................ 627
APPENDIX C
The ASCII Standard.......................................................................... 630
APPENDIX D
Numeric Limits and Precision .......................................................... 632
Accuracy and Precision.................................................................... 632
Significant Digits ............................................................................. 632
Computational Considerations ........................................................ 633
Quality Analyst Internal Numeric Data Types ................................. 634
APPENDIX E
References........................................................................................ 635
Citations .......................................................................................... 635
Where to Buy Books ........................................................................ 638
APPENDIX F
Equations.......................................................................................... 640
INDEX
Index ................................................................................................ 663
vi • Contents at a Glance
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
Table of Contents
CHAPTER 1
Introduction ..........................................................................................1
Overview of NWA Quality Analyst Features ........................................ 1
Data and SQC Chart Management ....................................................... 1
Graphics Presentation and Management ............................................ 1
General Statistical Analysis ................................................................. 2
Run File Interpreter............................................................................. 2
SQC Chart Server ................................................................................. 2
Background Knowledge Required ....................................................... 3
Statistical Quality Control ................................................................... 3
The Windows Environment.................................................................. 3
Typographical and Layout Conventions .............................................. 3
Technical Support................................................................................ 4
CHAPTER 2
Quality Analyst Files..............................................................................7
File Names ........................................................................................... 7
File Types and Their Extensions .......................................................... 7
Data Sets.............................................................................................. 8
Data Files ............................................................................................8
Header Files ........................................................................................9
Managing Data Sets ............................................................................9
Graphics Files..................................................................................... 12
Run Files ............................................................................................ 12
Summary Files ................................................................................... 12
Definition Files................................................................................... 13
Entering File Names........................................................................... 13
Registered File Types......................................................................... 14
Data File Structure ............................................................................ 15
Subgroup Types................................................................................. 17
Rowwise Subgroups .......................................................................... 18
Columnwise Subgroups ..................................................................... 19
Moving-Average Subgroups .............................................................. 21
Variable-Size Subgroups ................................................................... 23
Missing Data Handling....................................................................... 24
User-Defined Missing Data Symbols ................................................. 24
Missing-Data Techniques................................................................... 25
Tagged Data ...................................................................................... 26
CHAPTER 3
Creating and Editing Data Files...........................................................27
Opening an Existing Data File ........................................................... 28
Creating a New Data Set ................................................................... 29
Sorting Data....................................................................................... 98
Verifying Data .................................................................................. 100
Generating Random Numbers ......................................................... 101
Entering Data Using DDE ................................................................. 102
Dynamic Data Exchange—The Basics ..............................................102
How Quality Analyst Handles DDE Data ..........................................103
Configuring a DDE Connection ........................................................104
Customizing a DDE Connection .......................................................108
Closing a DDE Connection ...............................................................109
A DDE Example................................................................................109
Inserting Data from Another Data Set ............................................ 112
Saving a Data File............................................................................ 112
Printing Data ................................................................................... 113
Printing a Data Listing..................................................................... 113
Printing a Formatted Report ........................................................... 115
Selecting a Report Format ..............................................................115
Specifying a Report Format.............................................................116
CHAPTER 4
Selecting Variables........................................................................... 121
Selecting a Single Variable .............................................................. 121
Selecting Multiple Variables ............................................................ 123
Changing the Display Order ............................................................ 125
List By .............................................................................................. 125
Sort By Name ................................................................................... 125
CHAPTER 5
SQC Charting and Analysis ............................................................... 126
Selecting a Variable First................................................................. 126
Selecting a Chart Type First ............................................................ 127
Variable Control Charts ................................................................... 128
Creating Variable Control Charts..................................................... 128
Specifying Sample Size.................................................................... 129
Creating Cumulative Sum Charts .................................................... 130
Gallery of Variable Control Charts................................................... 131
Process Capability Charting and Analysis ....................................... 137
Process Capability Histogram.......................................................... 137
Process Capability Report................................................................ 137
Selecting Statistics and Display Codes ............................................139
Breakdown ......................................................................................139
Output Type ....................................................................................140
Attribute Control Charts .................................................................. 141
Pareto Diagrams.............................................................................. 144
Multiple Variables ............................................................................ 145
Single Variable................................................................................. 146
Analyzing Assignable Causes and Corrective Actions ..................... 148
Creating a Report of Assignable Causes and Corrective Actions .... 149
Charting Assignable Causes and Corrective Actions ....................... 152
Table of Contents • ix
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
Saving Assignable Causes and Corrective Actions to a Data File ... 155
Exception Reporting ........................................................................ 157
Creating an Exception Definition File.............................................. 158
Generating an Exception Report ..................................................... 160
Using an Exception Report .............................................................. 161
Viewing Exception Details...............................................................161
Charting Exception Data .................................................................162
Saving an Exception Report to a File...............................................163
Printing an Exception Report ..........................................................165
Other Charts .................................................................................... 166
Run Chart......................................................................................... 166
Scatter Diagram............................................................................... 169
Normal Probability Plot ................................................................... 171
Quantile Plot .................................................................................... 172
Weibull Analysis............................................................................... 174
Background .....................................................................................174
Weibull Distribution ........................................................................174
Exponential Distribution .................................................................175
Data Requirements .........................................................................175
Performing a Weibull Analysis ........................................................176
Using Weibull with Large Data Sets ................................................178
Box Plots .......................................................................................... 179
Selecting Variables..........................................................................180
Selecting Statistics ..........................................................................181
Breakdown ......................................................................................182
Box Type .........................................................................................182
Units ................................................................................................182
Creating Short Run Charts .............................................................. 183
Saving Charts .................................................................................. 183
Assigning Permanent File Names .................................................... 184
Viewing Graphics Files..................................................................... 185
Changing Your Mind ........................................................................ 185
Saving the Files................................................................................ 185
CHAPTER 6
Managing Charts with the Graphics Viewer .................................... 187
Running the Graphics Viewer Alone................................................ 188
Configuring the Graphics Viewer Interface..................................... 191
Individual Chart Windows............................................................... 192
Switching Between Chart Windows ................................................ 192
Arranging Multiple Chart Windows ................................................. 192
Minimizing Chart Windows.............................................................. 193
Printing Charts and Text Reports .................................................... 194
Configuring the Layout of Printed Charts and Reports................... 195
Single Charts or Text Reports ......................................................... 195
Group Layout ................................................................................... 196
Display the Charts ...........................................................................196
Create a Group Layout Window ......................................................197
Assemble the Layout .......................................................................198
x • Table of Contents
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
CHAPTER 7
Customizing Charts .......................................................................... 236
System Parameters ......................................................................... 242
General............................................................................................. 246
Title for Graphics and Reports.........................................................246
Date Format ....................................................................................248
Run File Error Handling ...................................................................248
Text Settings ...................................................................................249
Report Date.....................................................................................249
Table of Contents • xi
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
Decimal Symbol...............................................................................249
Time Delimiter.................................................................................249
Plotting Characters .........................................................................250
Limits and Specifications................................................................. 250
Limits and Specifications.................................................................251
Rainbow Zones................................................................................252
Chart Parameters............................................................................. 253
Calculation Methods for Charts .......................................................254
Basis of Control Limit Calculations..................................................255
Control Limit Adjustment for Subgroup Size...................................255
Pattern Rules ................................................................................... 256
Overview .........................................................................................256
Specifying Pattern Rules .................................................................258
Defining New Pattern Rules ............................................................259
Changing Pattern Rules...................................................................262
Deleting Pattern Rules ....................................................................262
Histogram Statistics ........................................................................ 262
Number of Footer Lines...................................................................263
Selecting Statistics ..........................................................................264
Capability Parameters ..................................................................... 266
Standard Deviation Calculation Method..........................................267
Capability Index Calculation Assumptions ......................................267
Sigma Factor ...................................................................................268
Confidence Limits............................................................................268
Calculated Bar Location ..................................................................268
File Paths and Permissions .............................................................. 269
Missing and Tagged Data ................................................................ 271
Tagged Data Mode ..........................................................................271
Missing Data Mode ..........................................................................272
Missing Data Symbol .......................................................................272
Box Plot and Capability Report ....................................................... 272
Capability Report Statistics.............................................................273
Box Plot Statistics ...........................................................................274
Box Plot Type ..................................................................................274
Box Plot Units..................................................................................274
External Database ........................................................................... 275
Assignable Cause Corrective Action ................................................ 275
Overview of Assignable Causes and Corrective Actions ..................275
Defining Lists of Assignable Causes and Corrective Actions ...........276
File Parameters ............................................................................... 278
File Title ........................................................................................... 278
File Range ........................................................................................ 279
Description Variables....................................................................... 279
X-Axis Description Variables ........................................................... 281
Maximum Variables on X Axis .........................................................281
Characters per Variable...................................................................281
Show Minor Tick Marks ...................................................................282
Display Interval...............................................................................282
Variable Parameters........................................................................ 283
Variable............................................................................................ 284
Data Ranges..................................................................................... 284
Confidence Limits............................................................................330
Capability Report Parameters ......................................................... 331
Default Variables ............................................................................. 331
Breakdown Defaults ........................................................................ 331
Box Plot Parameters........................................................................ 332
Default Variables ............................................................................. 333
Breakdown Defaults ........................................................................ 333
Pareto Parameters .......................................................................... 333
Pareto Format .................................................................................. 334
File Range (Multiple Variable Pareto only) ..................................... 335
Multiple Variable Pareto .................................................................. 335
Short Run Parameters ..................................................................... 336
CHAPTER 8
General Statistical Analysis.............................................................. 338
Overview ......................................................................................... 338
General Procedure ........................................................................... 339
Output of Statistics Routines .......................................................... 339
One-way ANOVA.............................................................................. 340
Completely Randomized .................................................................. 340
Randomized Block Design ............................................................... 342
Chi-square ....................................................................................... 344
Bartlett’s Chi-square Statistic ......................................................... 345
Difference Among Proportions ........................................................ 346
Equal Expected Frequencies ............................................................ 347
Unequal Expected Frequencies........................................................ 348
Correlation....................................................................................... 349
Kruskal-Wallis Statistic ................................................................... 351
Descriptive Statistics....................................................................... 353
t-Statistics ....................................................................................... 356
Ordinary Unpaired Sample t-Statistic ............................................. 356
Paired............................................................................................... 358
Single Sample .................................................................................. 360
Pearson Product Moment Correlation ............................................. 362
Wald-Wolfowitz Runs Test .............................................................. 364
Regression Analysis......................................................................... 366
Single-Variable Regression ............................................................. 366
Selecting Variables..........................................................................367
Regression Type ..............................................................................368
Graphics Output ..............................................................................369
Text Output .....................................................................................370
Confidence Limits............................................................................373
Multiple Regression ......................................................................... 374
Selecting Variables..........................................................................374
Graphics Output ..............................................................................375
Text Output .....................................................................................376
Confidence Limits............................................................................378
Polynomial Regression .................................................................... 378
Selecting Variables..........................................................................379
CHAPTER 9
Automating Procedures with Run Files............................................ 384
Creating Run Files ........................................................................... 385
How Run Files Work ........................................................................ 385
Run File Commands and Parameters .............................................. 386
General Form of Run File Commands .............................................. 391
Examples of Run File Commands ....................................................391
Requirements for Run File Commands ............................................392
ACCA ................................................................................................ 393
Text Report .....................................................................................393
Pareto Chart ....................................................................................394
Data File ..........................................................................................395
ANOVA.............................................................................................. 396
ANOVAR ........................................................................................... 397
APPEND............................................................................................ 397
AskAlpha AskDate AskInteger AskNumber AskTime ...................... 399
AskAlphaMatchFilter AskDateMatchFilter
AskIntegerMatchFilter AskNumberMatchFilter
AskTimeMatchFilter ......................................................................... 401
AskAlphaRange AskDateRange AskIntegerRange
AskNumberRange AskTimeRange ................................................... 403
AskAlphaRangeFilter AskDateRangeFilter
AskIntegerRangeFilter AskNumberRangeFilter
AskTimeRangeFilter......................................................................... 405
AskAlphaSelectFilter AskDateSelectFilter
AskIntegerSelectFilter AskNumberSelectFilter
AskTimeSelectFilter ......................................................................... 408
AskDATFile ....................................................................................... 411
AskList.............................................................................................. 413
AskOpenFile ..................................................................................... 416
AskSaveFile...................................................................................... 418
AskVariable ...................................................................................... 420
ATTRIBUT......................................................................................... 422
BOXPLOT .......................................................................................... 423
CALC................................................................................................. 425
CHART .............................................................................................. 426
CHITEST ........................................................................................... 427
CLEANUP .......................................................................................... 429
External to Quality Analyst—E.........................................................429
Fixed-Field to Quality Analyst—F.....................................................430
Quality Analyst to Comma-Delimited—S .........................................430
Quality Analyst to Comma-Delimited with Quotes—Q.....................430
Change Missing-Data Symbol—C.....................................................431
CLEAR............................................................................................... 431
Table of Contents • xv
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
CHAPTER 10
Using External Data Files ................................................................. 510
Importing Data................................................................................ 510
Importing a Comma-Separated-Value File...................................... 510
Importing a Fixed-Field File ............................................................ 512
Creating and Editing a Fixed-Field Definition File ...........................514
Opening a Data File without a Header File...................................... 518
Using an Existing Header File..........................................................519
Creating a New Header File.............................................................520
CHAPTER 11
Manipulating Data Files ................................................................... 530
Changing the Missing Data Symbol................................................. 530
Rotating a Data Set ......................................................................... 531
Converting Subgroups To or From a Single Column........................ 532
Single Column to Multiple Columns ................................................ 533
Multiple Columns to a Single Column.............................................. 535
Extracting Subsets from a Data File ................................................ 536
Extracting Data................................................................................ 536
Creating and Editing an Extract Definition File............................... 537
Defining Calculations for Run files .................................................. 538
Overview .......................................................................................... 538
Creating a Variable Definition File .................................................. 539
Defining Calculated Variables ......................................................... 540
Saving New Calculated Variables to a Data File.............................. 541
CHAPTER 12
Database Connectivity ..................................................................... 542
Overview ......................................................................................... 542
Before Using Connectivity ............................................................... 542
Direct Connection ............................................................................ 543
ODBC Connection............................................................................. 543
Defining Connectivity Data Sets...................................................... 543
Select New External Database ........................................................ 544
Selecting a Connection Type ........................................................... 545
Selecting a Data Source................................................................... 546
Direct-Connection Sources..............................................................546
ODBC Sources..................................................................................547
Specifying a Data Table ................................................................... 547
Using an Existing Table in the Data Source.....................................547
Denormalizing a Normalized Table .................................................548
Using SQL to Create a Data Table....................................................555
Setting External Database Access Options ..................................... 557
Assigning Database Fields to Quality Analyst Variables ................. 558
Selecting Fields Individually ...........................................................560
Selecting All Fields ..........................................................................561
Using Date-and-Time Fields ............................................................562
Defining Connectivity Data Filters .................................................. 563
Defining Filters to Retrieve Matching Data .....................................565
Defining Filters to Retrieve Similar Data.........................................566
Defining Filters to Retrieve Ranges of Data ....................................567
Using the (Ask) Settings..................................................................569
Removing Filters .............................................................................570
Specifying Key Columns ..................................................................570
APPENDIX A
Missing-Data Techniques in Control Charts..................................... 600
Missing Data with Rowwise Subgroups .......................................... 601
Mode 1 with Subgroup in Every Row .............................................. 602
Mode 2 with Subgroup in Every Row .............................................. 603
Mode 3 with Subgroup in Every Row .............................................. 604
Mode 4 with Subgroup in Every Row .............................................. 605
Mode 5 with Subgroup in Every Row .............................................. 606
Summary Overview—Rowwise Subgroups...................................... 606
Missing Data with Columnwise Subgroups ..................................... 606
Mode 1 with Subgroup Every 3 Rows.............................................. 607
Mode 2 with Subgroup Every 3 Rows.............................................. 609
Mode 3 with Subgroup Every 3 Rows.............................................. 610
APPENDIX B
Alternative Statistical Calculation Methods..................................... 624
EPA Control Limits........................................................................... 624
Ford Process Capability Indices ...................................................... 625
Control Chart Limits Based on Probability ...................................... 626
Standard Deviation Calculation Methods ........................................ 627
Variable Control Charts ................................................................... 628
Process Capability............................................................................ 628
Statistical Routines.......................................................................... 628
Practical Differences Between Methods.......................................... 629
APPENDIX C
The ASCII Standard.......................................................................... 630
APPENDIX D
Numeric Limits and Precision .......................................................... 632
Accuracy and Precision.................................................................... 632
Significant Digits ............................................................................. 632
Computational Considerations ........................................................ 633
Quality Analyst Internal Numeric Data Types ................................. 634
APPENDIX E
References........................................................................................ 635
Citations .......................................................................................... 635
Where to Buy Books ........................................................................ 638
APPENDIX F
Equations.......................................................................................... 640
INDEX
Index ................................................................................................ 663
xx • Table of Contents
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
CHAPTER 1
Introduction
NWA Quality Analyst is a Windows-based application that provides a
comprehensive, integrated set of Statistical Quality Control (SQC) charting and
analysis functions. Quality Analyst gives you the tools you need to analyze
quality-control data as well as produce the required control charts. It includes a
complete array of data-manipulation and -analysis capabilities. Quality Analyst
helps you locate and identify sources of variation, and can help you determine the
cause.
CHAPTER 1: Introduction • 1
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
annotate charts, view detail for individual data points, and customize the layout of
charts on the screen and printer. This routine can be run from the Editor or
independently to view and print graphs that have been saved from previous
Quality Analyst sessions or obtained from other users.
See CHAPTER 6: Managing Charts with the Graphics Viewer starting on
page 187.
You can also use Run files to integrate Quality Analyst functions with other
systems and applications.
See CHAPTER 9: Automating Procedures with Run Files on page 384.
2 • CHAPTER 1: Introduction
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
If you are not familiar with these terms, please refer to your Windows
documentation.
Though Quality Analyst adheres to most common conventions in the design of its
Windows interface, unique aspects of the Quality Analyst Windows user interface
are described throughout this manual as they occur.
CHAPTER 1: Introduction • 3
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
Typeface Description
<KEY STROKES> This font is used to indicate keys or combinations of keys that
you should press. For instance, <ENTER> indicates that you
press the Enter key, and <SHIFT+F1> indicates that you press
the <F1> key while holding down the <SHIFT> key.
On-Screen Text This font is used to indicate on-screen text such as menu items
to select, labels in dialog boxes, or buttons to click. For
example, the manual might say “Open the File menu and
select Open,” or “Click the OK button.”
Keyboard Entry Text that you enter using the keyboard is shown on its own
Text indented line in this typeface. For example:
Door Clearance (10th mm)
Also, when a button on the toolbar serves the same purpose as a menu selection,
that button will be shown in the margin to the left of the paragraph in which the
operation is explained. For example:
To create X-bar and Range charts of the data, open the Variable menu and select X-
Bar & Range.
Technical Support
Northwest Analytical is strongly committed to supporting the users of our
software. Registered users who have questions or encounter difficulties of any
kind with Quality Analyst are encouraged to contact us by phone, fax, or mail.
(Please call during normal business hours—8:30 PM–4:30 PM Pacific time.) An
experienced and interested staff is available to help, and there is no charge for this
service.
Our experience has shown that many support calls are the result of problems that
could have been solved by a careful re-reading of the instructions, or that were not
related to Quality Analyst in the first place. Before calling, please read (or re-read)
the pertinent section(s) of the manual, check input data for validity, and try to
reproduce the problem. If it is still necessary to call, please have the following
information available:
Your name and the name of your company or institution.
Product name (Quality Analyst), version, and serial number (required).
Type of computer and operating system.
4 • CHAPTER 1: Introduction
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
To allow us to help you effectively, please also have the following information:
Clear, concise description of the problem.
Source of the data involved.
Exact sequence of events (keystrokes, buttons, menu selections, and so on)
leading to the problem.
Description of program output.
Complete error message(s), if any.
It is always helpful, and sometimes essential, that you have access to your
computer system when calling. Although we can solve most problems quickly by
telephone, it may be necessary for you to provide printouts and/or the data files.
We encourage you to call with questions or problems. We are also interested in
comments and suggestions you have regarding Quality Analyst’s performance,
capabilities, ease of use, and documentation. We release updated versions of
Quality Analyst periodically, and user feedback is the largest single source of
ideas for these updates. If you have suggestions on how to improve Quality
Analyst, please let us know.
To contact NWA:
1 Open the Help menu and select Contact Northwest Analytical.
When you contact technical support, we may ask you to provide information about
your computer and your Quality Analyst installation.
CHAPTER 1: Introduction • 5
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
We may ask you to email us the information in this box by clicking the Email
button.
If you are troubleshooting database connectivity, we may also ask you to view the
connectivity debug log.
We may ask you to email us the information in this box by clicking the Email
button.
6 • CHAPTER 1: Introduction
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
CHAPTER 2
File Names
Quality Analyst file names can be any file name allowed by Windows. The
following are examples of valid file names:
SAMPLE.DAT VENEER.NWG Board Defects.dat
PART32 R12.DAT JM-42 Viscosity.NWH
1
ASCII stands for American Standard Code for Information Interchange.
Default
File Type Description
Extension
Data (*) DAT data only
Header (*) NWH definition of the data
Graphics NWG information needed to display a chart or text
report
Run RUN script for the “batch” processor
Summary SMY summary of out-of-control data points
Definition See page 13.
(*) A Data file and a Header file together constitute a Quality Analyst “Data Set.”
Data Sets
In Quality Analyst, the data and the information needed to describe the data and to
manage charting and analysis are stored in Quality Analyst “Data Sets.” A Data
Set consists of two files—a Data file that contains only data, and a corresponding
Header file that contains descriptions of the data and parameters used to analyze
and chart the data.
Data Files
Data files are created by you (using the Quality Analyst Editor) or obtained from
other sources. For example, Data files can be downloaded from remote computers
or can be created by other software. Data files normally contain the actual
measurements or data obtained directly from quality-control procedures. Data files
may also contain any number of descriptive fields to help identify individual data.
Data file names are assigned by you. The default file-name extension for Data
files is DAT. For example, a Data file might be named PARTS.DAT. For reasons
discussed below, under Managing Data Sets, we strongly suggest using this
default extension.
Quality Analyst Data files are space-delimited ASCII files in the form of a table.
But you probably won’t need to think about that unless you are moving files
between Quality Analyst and another software application. (See Data File
Structure below for more details.)
Normally, Quality Analyst will read and analyze all data in a Data file. But you
can also instruct Quality Analyst to read and analyze only some of your data. To
do this for an entire file, see the instructions in the File Range section under File
Parameters on page 279. To do so for individual variables within a file, refer to
Data Ranges under Variable Parameters on page 284. To do so for specific
combinations of chart type and individual variable within a file, refer to Center
Line and Limits under Chart Parameters on page 298. You can also apply filters
that will cause Quality Analyst to analyze only those rows with data that meet
certain criteria. This is explained under Filtering Data beginning on page 90.
Header Files
Header files, which are automatically created and maintained by Quality Analyst,
contain all the file-structure information needed to perform SQC charting and
analysis on the corresponding Data file. They includes the names and descriptions
of each variable in the Data file, the definitions of any calculated variables, and
parameters such as control limits and how they are calculated, chart scale and
labeling information, and product specifications used in process capability
analysis.
Header files have the same name as their corresponding Data files but use the
extension NWH. For example, the Header file for the Data file PARTS.DAT
would be named PARTS.NWH. Header files are created automatically by the
Quality Analyst Editor when you create Data files, and are updated as you make
changes to configuration parameters or the structure of the file. You can also
create and modify Header files from within Run files, as explained under Creating
Header Files from within Run Files on page 495.
Header file (for BOLTS.DAT) with the new Header file (for BOLTS.PRN) named
BOLTS.NWH. This will leave BOLTS.DAT without its Header file. Worse, if
you try to use BOLTS.DAT, Quality Analyst will use the Header for BOLTS.PRN
and an error will result.
For these reasons, we strongly suggest using the default Quality Analyst file-name
extensions for all Quality Analyst files. If you intend to use a file that was created
by another program but is usable “as is” by Quality Analyst, we suggest renaming
it first—before loading it into the Quality Analyst Editor or operating on it in a
Quality Analyst Run file.
To make managing Data Sets easier, Quality Analyst includes several file-
management commands, all of which appear on the Utilities submenu of the File
menu.
2 Specify the Data file you want to copy in the upper half of the dialog box.
3 If you want to copy both the Header file and the Data file, select the Copy
both .DAT and .NWH files check box. If you want to copy only one of the two
associated files, clear this check box and specify which file (.DAT or .NWH)
you want to copy.
4 Select a Destination folder in the lower half of the dialog box.
5 Click Copy.
3 If you want to delete both the Header file and the Data file, select the Delete
both .DAT and .NWH files check box. If you want to delete only one of the two
associated files, clear this check box and specify which file (.DAT or .NWH)
you want to delete.
4 Click Delete.
2 Specify the Data Set you want to rename in the upper part of the dialog box.
3 If you want to rename both the Header file and the Data file, select the
Rename both .DAT and .NWH files check box. If you want to rename only one
of the two associated files, clear this check box and specify which file (.DAT
or .NWH) you want to rename.
4 Select a new name for the Data Set in the New File Name box. You need not
supply a file name extension; Quality Analyst will apply the appropriate
extensions for each component of the Data Set.
5 Click Rename.
Note: You can also copy, delete, and rename files from within Run files. For
details, see COPY on page 432, DELETE on page 436, and RENAME
on page 466.
Graphics Files
Graphics files store information needed to display a chart or text report. They are
created by the Graphics Viewer and stored in XML format. Graphics files have the
extension NWG. Quality Analyst often creates several Graphics files from a single
Data file so it assigns temporary names like 1000001.NWG or 1.NWG (depending
on your license type) as it creates graphics output. The Editor lets you select
individual NWG files to save by renaming them.
Run Files
Run files are used in “batch processing.” Run files contain a series of commands
instructing Quality Analyst to perform its charting and analysis functions
automatically. Each row in a Run file contains the instructions necessary for
Quality Analyst to perform one of its functions. Run files are not necessarily
associated with a specific Data Set. They may contain instructions that let you
perform the same operations on different files with different parameters. You
create Run files using a text editor (like Windows Notepad) or a word processor
(using its ASCII-export or -save feature) or the built-in Run File Wizard. When
you name Run files, you must use the extension RUN.
Summary Files
Several Control Chart functions create Summary files that contain out-of-control
data points for display and printing. The default file name extension for Summary
files is SMY.
Definition Files
Definition files are special files used to store format information to increase the
efficiency of certain Quality Analyst functions. These functions and their
corresponding default file-name extensions are listed below. Quality Analyst
provides a user interface for creating and maintaining these files so you need not
be concerned with their structure. We recommend that you use the default
extensions for these file types to help keep track of these special files.
DEFECTS.DAT SAMPLE.NWG
HOUSING.DAT SAMPLE.RUN
If you want to use an existing file with an extension that is not the standard
Quality Analyst extension for that file type, you can first list those files by
entering the file name *.EXT (where EXT is the extension you want) and pressing
<ENTER>, then selecting your file from the list. Or you can enter the complete file
name, including the nonstandard extension. For example, if you created a Data file
named FAILRATE.PRN, you must enter FAILRATE.PRN when Quality Analyst
asks for the name of a Data file. If you created a file with no extension, you can
use it by typing a period at the end of the name. For example, if you created a Data
file named PART32, you must enter “PART32.” (with the period) when Quality
Analyst asks for the name of a Data file.
Remember, however, that we strongly suggest using Quality Analyst’s default
extensions for file names. See Managing Data Sets on page 9 for details.
When you import files to Quality Analyst from other programs, give them distinct
file-name extensions to prevent confusion, especially if you intend to run them
through the Quality Analyst conversion functions. For example, a “print file”
created in Lotus 1-2-3 will generally have the extension “.PRN” but does not need
converting. You should rename this file with the extension “.DAT” before using it
in Quality Analyst as a Data file. If you import and convert a file (such as a
comma-separated-value file, usually named with a “.CSV” extension) to Quality
Analyst format using one of the Import functions, you should name the output file
with the extension “.DAT.”
to one) in Explorer or on the Desktop, that file will be opened by the associated
Quality Analyst program component. This table shows which extensions are
registered with which component.
If you use other programs that also use these extensions, you can use Windows’
built-in features to unregister and re-register a file type. In Explorer, 2 open the
View menu, select Folder Options (or Options), then select the File Types tab.
Consult the Windows documentation for instructions on using this feature.
2
These menu selections apply to Windows 95/98 and may differ in subsequent versions.
separated from other values by spaces, data rows may not contain “blank” or
empty fields. Each field position must contain at least a “place holder.” For
example, if a file contains five variables, each line in the file must contain five
entries, separated by spaces. If a field datum is “missing,” its place in the row
must contain a missing-data symbol as a “place holder” for that data element to
maintain the structure of the Data file. The Quality Analyst Editor automatically
fills empty cells with the current missing data symbol when saving a Data file.
A typical Quality Analyst Data file, when loaded into a text editor, might look
something like this:
In this example, the first three columns are description fields (date, time, and
machine code), and the last four columns are the individual data elements of a
series of measurements (in this case, a subgroup). Each row is a separate sample—
in this example, a set of measurements. Both description and data elements are
separated by single spaces.
When loaded into the Quality Analyst Editor, this sample Data file (with
associated variable names and types supplied by you) looks like this:
The names at the top of each column are supplied by you as you define the data
structure when you create the Data Set (see Creating a New Data Set on page 29).
You use these names, which are stored in the Data Set’s Header file, to select
variables for analysis and to label graphs and reports. Each variable definition
includes a “data type,” which tells Quality Analyst what kind of information is
contained in the variable. Data types include DateTime, time, alphanumeric,
individual measurement, subgroup, and several others. Data types are discussed in
more detail under Data Type on page 37.
Measurements in Quality Analyst may be defined as individual values or as
subgroups—two or more related measurements. Subgroups may take many forms
(see the next section on Subgroup Types) but most commonly a single variable
spans multiple columns in the Data file.
Subgroup Types
Subgroups in Quality Analyst may be organized in one of four ways:
Rowwise Subgroups
Columnwise Subgroups
Moving-Average Subgroups
Variable-Size Subgroups
The type of subgroups allowed depends on the variable type of the data. (Quality
Analyst variable types and data types are explained in detail under Data Type on
page 37.) When the data is type “1, 2,…n” (subgroup), “X” (mean), “R” (range),
or “S” (standard deviation), each subgroup occupies one row of the Data file
(rowwise). With data type “I” (individual measurements), each subgroup can
occupy more than one row in a Data file (columwise). When the subgroups are
processed as moving averages (or moving ranges, or moving standard deviations),
consecutive subgroups “share” data rows.
The following sections describe each type of subgroup and provide an example of
each.
Rowwise Subgroups
In “rowwise” subgroups, the measurements for a subgroup variable are all in one
row of the Data file. A subgroup variable of five measurements would occupy five
fields in one row of the Data file. That is, each Data file row contains a complete
subgroup.
In this example, the subgroup size is three and each row contains the three
measurements for one subgroup. The X-bar chart values show how subgroup
values are combined for charting and analysis.
Columnwise Subgroups
In “columnwise” subgroups, each row of a Data file contains only one
measurement of a subgroup variable. Therefore, each subgroup occupies several
rows of a Data file. One subgroup begins on the row immediately following the
end of the previous subgroup. For example, with a columnwise subgroup of five
measurements, each subgroup would occupy five Data file rows, and a new
subgroup would begin every five rows.
In this example, the subgroup size is three, but the data all is located in one
column. Each subgroup occupies three rows of the Data file, and a subgroup
begins every three rows. The X-bar chart values show how subgroup values are
combined for charting and analysis.
Data File Row Data File Contents Subgroup Number X-bar Chart Values
1 5.5 1
2 5.6 1
3 7.3 1 6.13
4 7.1 2
5 6.7 2
6 6.8 2 6.87
7 5.4 3
8 6.4 3
9 5.4 3 5.73
10 5.6 4
11 5.3 4
12 6.9 4 5.93
Moving-Average Subgroups
In “moving-average” subgroups, each row of the Data file contains only one
measurement of the subgroup variable. Therefore, each subgroup occupies several
rows of a Data file. Unlike columnwise subgroups, however, a new subgroup
begins on each row of the Data file. For example, with a subgroup of five
measurements, each subgroup would use five rows of a Data file, and a new
subgroup would begin at each row. This type of subgrouping produces moving
average charts.
In this example, the subgroup size is three, but the data all is located in one
column. Each subgroup requires three Data file rows, but a subgroup begins on
every row. The X-bar chart values show how subgroup values are combined for
charting and analysis.
Variable-Size Subgroups
In “variable-size” subgroups, each row of the Data file contains only one
measurement of a subgroup variable. Therefore, each subgroup occupies several
rows of a Data file. The subgroups vary in size, and are identified by the value of
an associated variable. A subgroup begins on the row immediately following the
end of the previous subgroup. For example, if a subgroup consists of a single
day’s samples, a variable such as DATE would be used to identify each subgroup;
a new subgroup begins each time the value of DATE changes.
In this example, a new subgroup starts each time the DATE variable changes. The
actual data occupy only one column. Each subgroup occupies a different number
of rows in the Data file. The X-bar chart values show how subgroup values are
combined for charting and analysis.
Notice in the figure that Quality Analyst automatically adjusts the control limits
for differences in subgroup size.
Missing-Data Techniques
In Quality Analyst control charts, you have several options for handing missing
data, as discussed in APPENDIX A: Missing-Data Techniques in
Control Charts.
Most Quality Analyst statistical functions deal with missing data by ignoring cases
with missing data and adjusting the results accordingly.
Calculated variables that use formulas that include missing data may result in the
missing-data symbol depending on the functions and operators in the formula.
The following functions accept multiple arguments (including ranges of variables)
and will result in numeric values if some (but not all) of the arguments are missing
data. The missing-data arguments are simply ignored.
AVG MIN STDDEV
MAX RANGE SUM
The following functions accepts only a single argument and will result in missing
data if the argument is missing data:
ABS GRD SIN
ATN INT SQR
COS LN TAN
DEG LOGn
EXP MOD
Here are some examples of functions operating on missing data (represented
by [MD]):
AVG(4,[MD],7,6) = 5.6666667
MAX([MD],-1.5,2.6,3.0) = 3.0
MIN([MD],[MD],[MD]) = [MD]
COS([MD]) = [MD]
The addition (+) and subtraction (−) operators produce numeric results if at least
one term is not the missing-data symbol. The multiplication (*) and division (/)
operators will result in missing data:
4.3 + [MD] = 4.3
-4.3 − [MD] = -4.3
Tagged Data
Quality Analyst lets you “tag” data for “special” treatment. Quality Analyst can
treat tagged data either as normal data or as missing data, according to the current
settings for handling missing data, described in APPENDIX A. By default,
Quality Analyst treats tagged data as missing. Tagging data and treating it as
missing is primarily done to eliminate selected values from calculations of control
limits.
You can configure the way Quality Analyst treats tagged data globally, as
explained under Missing and Tagged Data on page 271.
You can override that global setting for each variable in a Data file, as
explained under Tagged Data Handling on page 287.
How to tag data is explained under Tagging Data on page 81.
CHAPTER 3
The Quality Analyst start-up screen, shown below, includes a menu, toolbar, and a
console with prominent buttons for common start-up operations:
Create a new Data Set
Open an existing Data Set
Execute Run files
Settings
Help
This is the standard Windows dialog box, except that, when you select a Data file
in the file list, the Quality Analyst file Title (if the file contains one) appears at the
bottom of the dialog box.
Before you can enter data into a Quality Analyst Data file, you must define the
structure of the data. In this dialog box, you choose either to define that structure
“from scratch” (using the Define New button) or to use the structure of an existing
file as a starting point (using the Use Existing button).
If you click Use Existing, the Quality Analyst Open dialog box (described under
Opening an Existing Data File above) will be displayed. Select the file whose
structure you want to borrow for your new file. The main Editor data grid will
appear with column headings representing the data structure of the file you
selected.
If you click Define New, an empty Variable Definition and Specifications form
(shown below) will appear. The upper portion of this dialog box contains a
spreadsheet-like grid in which you define the structure of your Data Set. Before
we explain the use of this form, we will describe the editing functions and features
of the spreadsheet-like grid.
If you want to create a connectivity Data Set, click New External Database. Refer
to CHAPTER 12: Database Connectivity on page 542 for additional instructions.
Note: The editing functions of the main data-set grid of the Quality Analyst
Editor are slightly different (and more numerous) and are described under
Editing Functions on page 73.
User Action or
Result
Keystroke
<RIGHT> Selects the cell to the right of the currently selected cell.
<LEFT> Selects the cell to the left of the currently selected cell.
<UP> Selects the cell above the current cell.
<DOWN> Selects the cell below the current cell.
<ENTER> Selects the cell to the right. Skips from the last cell in the row
to the first cell in the next row.
<TAB> Same as <ENTER>.
<SHIFT+TAB> Selects the cell to the left. Skips from the first cell in the row to
the last cell in the previous row, if any.
<HOME> Selects the cell in the first column of the current row.
<END> Selects the last nonblank cell in the current row.
<CTRL+HOME> Selects the cell in the first column of the first row.
<CTRL+END> Selects the last nonblank cell in the last row.
<PGDN> Selects the cell one screenful of rows below the current cell.
<PGUP> Selects the cell one screenful of rows above the current cell.
Mouse Click Selects the cell at the mouse pointer.
User Action or
Result
Keystroke
<ESC> Enters edit mode and selects the current contents of the cell.
<DELETE> Same as <ESC>.
Double-Click Same as <ESC>.
<DELETE><DELETE> Deletes current cell contents and enters edit mode.
<ESC><RIGHT> Enters edit mode with insertion point at end of current cell
contents.
<ESC><HOME> Enters edit mode with insertion point at beginning of
current contents.
Text Enters edit mode and replaces current cell contents with
“Text”.
These actions place and leave you in edit mode. While in edit mode, the thick
black border around the cell will disappear and the Windows text-insertion cursor
(normally a vertical blinking bar) will appear. The following keystrokes apply
while you are in edit mode.
User Action or
Result
Keystroke
<LEFT> Moves insertion point one character to the left.
<RIGHT> Moves insertion point one character to the right.
<HOME> Moves insertion point to the left of all text.
<END> Moves insertion point to the right of all text.
<UP> In a cell with a drop-down button, selects the previous option in
the list. Otherwise, accepts current cell contents and selects the
cell above.
<DOWN> In a cell with a drop-down button, selects the next option in the
list. Otherwise, accepts current cell contents and selects the cell
below.
<ENTER> Accepts current cell contents and selects the next cell (skipping
from end of row to next row).
<ESC> Restores original value and exits edit mode.
<TAB> Same as <ENTER>.
<SHIFT+TAB> Accepts current cell contents, exits edit mode, and selects the
previous cell.
<CTRL+RIGHT> Accepts current cell contents and selects cell to the right.
<CTRL+LEFT> Accepts current cell contents and selects cell to the left.
<DELETE> If any text is selected, deletes selected text. Otherwise, deletes
character to the right of the insertion point.
<BACKSPACE> If any text is selected, deletes selected text. Otherwise, deletes
character to the left of the insertion point.
<SHIFT+> When used with <LEFT>, <RIGHT>, <HOME>, or <END>,
selects text in the direction of the move.
Mouse Click In same cell, moves insertion point to under the mouse pointer
(I-beam mouse cursor).
Otherwise, accepts current contents of the cell and selects the
cell under the mouse pointer (fat “+” cursor).
Text Enters “Text” into the cell.
The cells in the Type column in both forms each contain a drop-down list. You can
select a data type using any of the standard Windows methods for selecting an
item from a drop-down list.
To delete a variable in a data grid, select it by clicking its row number at the left of
the form, then click either the Delete or Cut buttons. If you click Delete, a warning
dialog box will appear asking you to confirm the deletion of the variable.
Click OK to delete the variable. If you click Cut, the variable definition will simply
disappear but a copy of it will be stored in the Clipboard for possible retrieval. In
both cases, the variables below the deleted variable will move up to fill the gap.
To move a variable in the grid to a new position, first Cut its definition to the
Clipboard as just explained. Then select the variable above which you want to
insert the cut variable. Click the Insert button to insert a blank row in the form,
then click the Paste button to paste the variable definition from the Clipboard.
first column in the data-entry grid is unlabeled and simply numbers each variable.
The remaining columns are labeled Name, Type, Description, LSL, USL, and Target.
The Variable Parameters tab contains additional variable-definition parameters. If
you are creating your data structure from scratch, you should fill in this table one
row at a time, from left to right. If you are modifying a data structure borrowed
from another file, edit this table if necessary.
Note: Detailed explanations of the editing functions and features of this form
are presented under Data Structure Editing Functions on page 31.
Variable Name
In the column labeled Name, you enter a name for each variable. Variable names
have the following restrictions:
They must be 2 to 32 characters long.
They may contain letters (“A”–“Z”), numerals (“0”– “9”), or the underscore
(“_”).
They must begin with a letter.
Each must be unique within a Data file.
Data Type
After you name the variable, you must indicate what type of variable it is. In the
column labeled Type, you can either type in a legal Quality Analyst variable type,
or you can click the button in the right end of the cell and select the variable type
from the resulting drop-down list. Quality Analyst variable types can be composed
of one or more data types depending on the type of data, the form they are in, and
the analyses you intend to perform on them. Variables composed of more than one
data type have a single name but are displayed in the Editor in multiple columns.
The following table lists all legal Quality Analyst data types. The table below that
lists all legal variable types (showing how the data types can be combined) and
notes their common application.
Reduced Data
“Reduced” data is data that is represented by the mean, range, and standard
deviation of measurements in a subgroup. This type of data is often available
instead of raw subgroup measurements, possibly from a measuring system that
performs the calculations on the measurements before storing or transmitting the
data. The variable types are designated X (X-bar or mean), R (range), and S
(standard deviation). The data for a single variable should have the same variable
name. For example:
DateTime
DateTime variables are designated type D. The value of a DateTime variable must
include a date and may include a time, separated from the date by a space.
Quality Analyst recognizes three date formats.
Format Example
Month/Day/Year 10/2/06
Day/Month/Year (European) 2/10/06
Day/Month/Year (Military) 2-OCT-06
All dates must contain the year which can be either two or four digits, for
example: 06 or 2006. Date formats may not be mixed within a file.
Date delimiters must be either “/”, “-”, or “.”, and must not be mixed within a file.
Do not use spaces or commas as date delimiters, since Quality Analyst would treat
each element as a separate field.
The first two example date formats require a numeric representation for the month.
For the military date, the months must be represented by JAN, FEB, MAR, and so
on. The characters may be all uppercase, all lowercase, or “proper-name” case
(one uppercase letter and two lowercase letters), for example JAN, jan, or Jan.
You should use case consistently, since many Quality Analyst functions
distinguish between the same character in a different case.
The date format in the Data file must be the same as that specified globally as
explained under Date Format on page 248.
If the value of a DateTime variable includes a time, enter it in the format HH:MM
or HH:MM:SS where HH is the hour (00–24), MM is minutes (00–59), and SS is
seconds (00–59). If the time element is absent, Quality Analyst assumes it is
00:00:00 (midnight at the beginning of the day). If the seconds element is absent
in a time element, Quality Analyst assumes it is 00.
Leading zeros in any date or time element are acceptable but not necessary. For
example, the date 12-01-06 is treated the same as 12-1-06, and the time 03:05 is
treated the same as 3:05 and 3:5.
Time
Time variables are designated type T and contain elements for hour, minutes, and
seconds. Time delimiters may be either “.” or “:” (to configure, see Time
Delimiter on page 249) and must not be mixed in the same file. Leading zeros are
unnecessary.
Format Example
Hour:Minute 14:20
Hour.Minute 6.30
Hour:Minute:Seconds 18:37:32
If you intend to sort or filter your Data file by time, you should use the 24-hour
time format.
If you are using an existing file that contains a time field in a 12-hour format and
you intend to sort or filter the file by time, you should add an alphabetic
description field (see below) to your file to indicate whether the time is A.M., P.M.,
noon, or midnight. In this field, enter “AM” for A.M., “PM” for P.M., “NN” for
noon, and “AA” for midnight. If you have configured Quality Analyst to be case-
sensitive when sorting or filtering on alphabetic fields (as explained under Text
Settings on page 249), be sure to use case consistently, since “PM” and “pm” are
treated differently when Quality Analyst is configured that way. (By default,
Quality Analyst ignores case in alphabetic comparisons and considers “PM” and
“pm” to be identical.)
The time format in the Data file must be the same as that specified globally as
explained under Time Delimiter on page 249.
PART4(G) PART4(M)
23.5 5
29.0 8
27.1 3
TEST(C) TEST(M)
BENT 15
SCRATCH 12
DIRTY 13
Other Data
Data that does not match one of the other types is designated type O. Use this type
for numeric data that is neither measurements nor defects, but may be analyzed by
one of the statistical functions.
Description
In the column labeled Description, enter a short description of what the variable is.
This will be used for labeling charts and reports. Descriptions may be up to 40
characters long (including spaces and punctuation). For measurement variables,
you might want to include the units of the measurement, for example,
Door Clearance (mm)
Note: If you are not reading specifications from a database, you can skip to
Other Variable-Definition Parameters on page 55.
You can configure Quality Analyst to read specifications from a database. Quality
Analyst can read specifications stored in either of the two general table formats
shown in the following examples.
In the format above, each row of the data table contains all specifications, in
separate columns, for a single attribute (for example, length or weight) of a single
product. Other columns indicate the product and attribute.
In the format above, each row of the data table contains a single specification and
the columns indicate the product, measurement name, specification type, and
specification value.
If your specifications are stored in a table that does not match either of these
general formats, you can create your own SQL query to read specifications.
To start the process, click the External LSL, USL, Target button in the Variable
Definition and Specifications dialog box. The External Specification Database Setup
dialog box will appear.
In the Connection list, select the type of database in which your specifications are
stored. The list contains most popular database formats.
In most cases, when you select a Connection type, a corresponding Open dialog
box will appear.
Select the source where your specifications are stored, then click Open. The name
of the source will appear in the Data Source box.
If you specify ODBC for Connection, no Open dialog box will appear. Instead, the
Data Source box becomes a drop-down list that contains all the Data Source
Names (DSNs) defined on your system. Open the list and select the DSN of the
database that contains your specifications.
If you need to create a custom SQL query to retrieve specifications, you can skip
to Entering a Custom SQL Query on page 54.
In the USL Field list, select the database field (column) that contains the upper
specification limits for the measurements. In the example table, it is “USL.”
In the LSL Field list, select the database field (column) that contains the lower
specification limits for the measurements. In the example table, it is “LSL.”
In the Target Field list, select the database field (column) that contains the target
values for the measurements. In the example table, it is “Target.”
In the Limit Value Field list, select the database field (column) that contains the
values of the specifications. In the example table, it is “SpecValue.”
In the Limit ID Field list, select the database field (column) that contains the type of
the specification. In the example table, it is “SpecType.”
In the USL Name box, enter the value of the Limit Value Field that corresponds to
the upper specification limit for the measurements. In the example table, it is
“USL.”
In the LSL Name box, enter the value of the Limit Value Field that corresponds to
the lower specification limit for the measurements. In the example table, it is
“LSL.”
In the Target Name box, enter the value of the Limit Value Field that corresponds to
the target for the measurements. In the example table, it is “Target.”
Select the database field (column) whose values you want to filter by, then click
OK. That field will appear in the Database Field(s) list.
In the editable Must Match list, specify the value of that database field for which
you want to read specifications. Do one of the following:
Enter a legal value of that database field.
Select (Ask) from the drop-down list. Each time Quality Analyst queries the
database for specifications, Quality Analyst will prompt you to enter a value
for the database field.
Select one of the field names, if any, in the list. If your Data Set is configured
to connect to an external database, and it is configured to prompt you, when it
connects, for the value of one or more database fields, then those database
fields will appear in this list. If you select one of those database fields here,
then the value you enter when Quality Analyst connects to the database will
be used for this specifications filter.
To remove a filter, select the field in the Database Field(s) list, then click Remove.
2 Verify that the correct specifications are being retrieved for the corresponding
variables.
3 Click OK.
When your settings produce the desired results, click OK in the External
Specification Database Setup dialog box to return to the Variable Definition and
Specifications dialog box.
3 Select the value of the Variable Name Identifier field (from the Database
Specification Setup dialog box) that corresponds to the selected variable.
4 Click OK.
In the External Specification Database Setup dialog box, select User Supplied SQL,
then click the adjacent Enter SQL button. The User Supplied SQL dialog box will
appear.
In the USL, LSL, and Target boxes, enter (or paste from another application) the
SQL commands that will retrieve the appropriate values for each specification
respectively.
Select the Pass SQL to Database for processing check box if you want Quality
Analyst to send the SQL query directly to the database to be processed. You will
want to do this if your query uses the “native” SQL dialect of your database. If
you clear this check box, Quality Analyst will send your query to the Microsoft Jet
database engine on your system, in which case you should use Microsoft
Jet/Access SQL syntax.
Click OK.
The area labeled Parameters will display the current settings for the selected
variable. You can select a variable by clicking its name in the Variables list at the
left of the Variable Parameters tab, or by selecting any cell in the row for the
variable in the grid above.
Input Mask
In the box labeled Input Mask, you can enter a “template” or “mask” to make
entering data easier. The mask lets you avoid entering unchanging or repetitive
portions of input for every datum. A mask can contain “replaceable” and “fixed”
characters. Replaceable characters in the mask will be replaced by the characters
you enter into a data cell to which the mask applies. Fixed characters will be
added to the characters you enter.
There are two allowable replaceable characters—the pound sign (#) and the
ampersand (&).
Use # to allow only digits to be entered with an input mask.
Use & to allow any character to be entered with an input mask.
Any other characters in the input mask are treated as fixed characters. The
following table shows some sample input masks and their behavior with sample
entries.
While editing the grid, when you select a cell to which an input mask applies, the
status line at the bottom of the Quality Analyst window displays Input Mask:
followed by the mask. You can override the input mask temporarily by pressing
<CTRL+O> or by opening the Edit menu and selecting Override Input Mask. This
is described in more detail under Overriding an Input Mask on page 87.
Column Width
In the box labeled Width, you can specify the approximate number of characters
for the width of the data column when displayed by the Editor. The units of Width
are the approximate number of characters since the widths of characters in most
Windows fonts are not uniform. This setting affects only the display of the values
in the Editor and does not cause data values to be rounded or truncated.
Note: You can also set the column width in the Editor using your mouse by
dragging the right edge of the column header left or right.
Decimal Places
In the box labeled Decimal Places, enter the number of decimal places for the
variable. (This setting is available for numeric data only.) As you enter data for
this variable in the Editor, Quality Analyst will round it off to this many decimal
places before placing it in the cell.
Note: This setting affects not just the display of the data but the data
themselves. Only the number of decimal places you specify will be stored
in the Data file.
Axis Label
Quality Analyst will use the text you enter in the box labeled Axis Label to label
the X or Y axis of graphs when the selected variable is plotted along that axis.
Normally, Quality Analyst will use either the variable name or no label. This
setting lets you override that behavior and specify a label that is longer, shorter, or
perhaps more “readable” than the variable name.
Specification ID
This setting is used to assign specifications that are read from a database to the
appropriate Data Set column. Its use is explained under Assigning Specifications
to Variables on page 52.
uniquely identify each record in the Data Set. Usually, a date and a time field will
suffice, but you might need to include a lot or part number as well.
You must arrange the key columns in order of decreasing significance—that is,
most significant (coarse) first, and least significant (fine) last.
The values of the key columns are used to associate the data in the Data Set with:
Graphics comments assigned to charted data points (see Annotating Charts
on page 203)
Assignable causes and corrective actions assigned to charted data points (see
Assigning Causes and Corrective Actions to Data Points on page 217)
Limit regions—regions of data identified for the calculation or application of
control limits (see Defining Regions on page 304)
Tagged data
We strongly recommend that you identify key columns for your Data Sets in all
cases, but if you are certain you will never use any of the features listed above,
you can forego selecting any key columns.
Note: If you are defining a Connectivity Data Set, you should always select key
columns for your Data Set. You can do so either as explained above, or
in the Database Connectivity Definition dialog box as explained under
Specifying Key Columns on page 570.
Note: Before you can assign lists of causes and corrective actions to a variable
on this tab, you must have already created those lists. See Assignable
Cause Corrective Action on page 275 for instructions.
In that section, you may also find Overview of Assignable Causes and
Corrective Actions helpful.
If you are satisfied with the display order of the data columns and do not need to
define any calculated variables, the settings in this form are optional and you may
ignore this section.
To perform either of these operations, open the Edit menu and select Column
Definition and Calculations. The Column Definition and Calculations dialog box will
appear.
Each row of this grid corresponds to a column of data as displayed in the Editor.
The first column of the form is unlabeled and simply numbers each row of the grid
(or column of data). The remaining columns are labeled Name, Type, =, and
Calculation Definition.
Note: Detailed explanations of the editing functions and features of this form
are presented under Data Structure Editing Functions on page 31.
Note: The variables created by this procedure contain live data. That is, if the
values of any variables in the formula change while you are editing the
file, the value of the calculated variable will also change (as in a
spreadsheet). If instead, you want to calculate values based on other
variables and insert them as fixed values into a new data column, see
Saving New Calculated Variables to a Data File on page 541.
Define calculated variables in the Column Definition and Calculations dialog box
which you can display by opening the Edit menu and selecting Column Definition
and Calculations. Calculations are defined using the columns labeled = and
Calculation Definition.
In all cases, the cell in the Calculation Definition column will become a large gray
text box in which you compose your formula. You can type the formula from the
keyboard or use the Calculation Parameters area that will appear in the lower half
of the dialog box. See Composing a Formula below.
Column Definition and Calculations dialog box with formula composition box
awaiting a formula
When you have composed your formula, click OK in the Definition Parameters
area. Quality Analyst will check the formula for errors. See Correcting Errors in
Formulas below.
Calculation Parameters
You can use the Calculation Parameters area of the dialog box to compose a
formula by pointing and clicking variable names and operators.
Composing a Formula
Formulas for calculating variables follow common algebraic rules for notation and
order of evaluation. They may include any variable name in the input file and any
of several common mathematical functions or operations.
Variable Names
To base the calculation on an existing variable, include its name and data type,
separated by a colon, in the formula. For example:
PITTED:I + BUBBLES:I + MARKS:I + EDGE:I
If the variable comprises subgrouped measurements (type 1, 2, 3, and so on),
replace the data type with the subgroup element number. For example:
(DIAM:1 + DIAM:2 + DIAM:3) / 3
Operators
The following table lists the operators and functions available for composing
formulas for calculated variables. Enclose arguments to functions and operators in
parentheses. For example:
LN(SIZE:I)
Multi-argument Functions
The multi-argument functions operate on two or more arguments. These
arguments may be passed as individual arguments or a range of columns.
Separate individual arguments with commas:
MIN(LENGTH:I,WIDTH:I,HEIGHT:I)
MAX(WEIGHT:I,550)
To pass a range of columns as an argument, separate the first column from the last
column with a tilde (~):
RANGE(WEIGHT:1~WEIGHT:5)
Order of Evaluation
The elements of a formula are calculated in the following order:
1 functions
2 expressions within parentheses
3 exponentiation
4 multiplication and division
5 addition and subtraction
For example:
SIN(SIZE:I)^2 [the square of the sine of SIZE]
SIN(SIZE:I^2) [the sine of the square of SIZE]
Use parentheses to override normal evaluation order:
2*(EDGE:I+CENTER:I) [add first, then multiply]
Function Modifiers
For logarithms other than natural logarithms (base e), include two arguments in
parentheses—the number to take the log of and the base of the logarithm function.
For example:
LOGn(WEIGHT:I,10) [base-10 logarithm of WEIGHT]
Trigonometric functions are performed in radians. Convert values in degrees or
grads to radians before passing them to trigonometric functions using one of the
unit-conversion functions—DEG or GRD. For example:
SIN(DEG(LENGTH:I)) + COS(DEG(DIAM:I))
SIN(GRD(LENGTH:I)) + COS(GRD(DIAM:I))
Polynomial DIAM cubed plus 3.2 times DIAM squared minus 1.32 times DIAM
minus 1.0:
DIAM:I^3 + 3.2*DIAM:I^2 - 1.32*DIAM:I - 1.0
Function 6.3 plus 1.25 times LENGTH minus 0.89 times WIDTH plus 22.5 times
HEIGHT:
6.3 + 1.25*LENGTH:I - 0.89*WIDTH:I + 22.5*HEIGHT:I
Natural log of LENGTH plus the sine of WIDTH (in radians) minus the reciprocal
of the difference between HEIGHT and 1.0:
LN(LENGTH:I) + SIN(WIDTH:I) - 1/(HEIGHT:I-1)
The square of DIAM plus e raised to the power THICK:
DIAM:I^2 + EXP(THICK:I)
The date portion of the DateTime variable TIMESTAMP:
DATEONLY(TIMESTAMP:D)
You can change the order in which data columns are displayed in the Editor by
rearranging the rows in this grid. The section Inserting, Deleting, and Moving
Variables on page 35 describes how to rearrange rows in the grid.
Another word of caution: Deleting a column using either method deletes the
corresponding data from your Data Set. If you do this inadvertently, recover the
data by exiting the Editor without saving the changes, then start the Editor again
and reopen the Data file. If you try to delete a column that is used to calculate
another column, Quality Analyst will warn you that the formula for the calculated
column will be deleted, and ask you to confirm your intention to do this. If you
confirm, the values in the calculated column will be converted to fixed values and
inserted into the column before the formula is deleted.
Editing Functions
This section describes the data-editing functions and features available in the
Quality Analyst Editor. (If items on the Edit menu are disabled, see File Paths and
Permissions on page 269 or contact your system administrator.)
3 Click OK.
2 Select or clear the check box next to each button or separator to include or
exclude it from the toolbar.
3 Click OK.
Note: The functions of all toolbar buttons are available from Quality Analyst’s
menus, so you will not lose functionality by removing a button from the
toolbar.
Quality Analyst provides several ways of moving from cell to cell. These are
summarized in the table below.
User Action or
Result
Keystroke
<RIGHT> Selects the cell to the right of the currently selected cell.
<LEFT> Selects the cell to the left of the currently selected cell.
<UP> Selects the cell above the current cell.
<DOWN> Selects the cell below the current cell.
<ENTER> Moves to the next cell. In Row Mode, skips from the end of
a row to the first column in the next row, if possible.
<TAB> Same as <ENTER>.
<SHIFT+TAB> Moves to the previous cell. In Row Mode, skips from the
beginning of a row to the last column in the previous row, if
possible.
<HOME> Selects the cell in the first column of the current row.
<END> Selects the cell in the last column of the current row.
<CTRL+HOME> Selects the cell in the first column of the first row.
<CTRL+END> Selects the cell in the last column of the last row.
<PGDN> Selects the cell one screenful of rows below the current cell.
<PGUP> Selects the cell one screenful of rows above the current cell.
Click Selects the cell at the mouse pointer.
in—row or column. The current cursor mode is shown in the lower right-hand
corner of the Editor where either Mode: ROW (the default) or Mode: COLUMN will
be displayed. 3 (Both the Variable Definition and Specifications and Column
Definition and Calculations dialog boxes are always in Row Mode, so the mode is
not displayed.)
In Row Mode, the “next” cell is the first cell to the right of the current cell. If the
current cell is the last in the row, the next cell is the first cell in the row just below
the current row. The previous cell, in Row Mode, is the cell to the left of the
current cell. If the current cell is the first in a row, the previous cell is the last cell
in the row just above the current row.
In Column Mode, the “next” cell is the cell below the current cell, and the
previous cell is the cell above the current cell, if any. If the current cell is in the
last row of data, the next cell is not the top cell in the column to the right. This lets
you add new rows of data using keystrokes that move to the next cell.
Note: Quality Analyst will never let the “next” cell be in a calculated column.
2 In the Go To a Data Cell dialog box, enter the row number and column number
of the cell you want to select.
3
Both the Variable Definition and Specifications and Column Definition and Calculations dialog
boxes are always in Row Mode, so the mode is not displayed.
3 Click OK.
Modifying Data
To modify the contents of the selected cell, use the following keystrokes or mouse
actions.
User Action or
Result
Keystroke
<INSERT> Enters edit mode and places insertion point at end of entry.
Input mask, if any, is not enforced.
<ESC> Same as <INSERT>.
Double-Click Same as <INSERT>.
Text Enters edit mode and replaces current cell contents with
“Text”. Input mask, if any, is displayed and enforced.
<DELETE> Deletes cell contents and enters edit mode. Input mask, if
any, is displayed and enforced.
<CTRL+O> Same as <DELETE> except Input mask is ignored.
These actions place and leave you in edit mode. While in edit mode, the cell will
turn gray and will have a thin black border and the Windows text-insertion cursor
(normally a vertical blinking bar) will appear. While you are in edit mode the
following keystrokes apply.
Edit-Mode Keystrokes
User Action or
Result
Keystroke
<LEFT> Moves insertion point one character to the left.
<RIGHT> Moves insertion point one character to the right.
<HOME> Moves insertion point to the left of all text.
<END> Moves insertion point to the right of all text.
<UP> Accepts current cell contents and selects the cell above.
<DOWN> Accepts current cell contents and selects the cell below.
<ENTER> Accepts current cell contents and selects the next cell.
<ESC> Restores original value and exits edit mode.
<TAB> Accepts current cell contents and exits edit mode.
<CTRL+RIGHT> Same as <END>.
<CTRL+LEFT> Same as <HOME>.
<SHIFT+TAB> Same as <TAB>.
<DELETE> If any text is selected, deletes selected text. Otherwise,
deletes character to the right of the insertion point.
<BACKSPACE> If any text is selected, deletes selected text. Otherwise,
deletes character to the left of the insertion point.
<SHIFT+> When used with <LEFT>, <RIGHT>, <HOME>, or <END>,
selects text in the direction of the move.
Click In same cell, moves insertion point to under the mouse
pointer (I-beam mouse cursor).
Otherwise, accepts current contents of the cell and selects
the cell under the mouse pointer (fat “+” cursor).
Text Enters “Text” into the cell.
In the box labeled Insert Row(s), enter the number of rows you want to insert.
The Insert Row(s) box will display the number of rows you selected before
invoking the dialog box. Change this value if you want to insert a different
number of rows. The Beginning at Row box will display the number of the
row you selected before invoking the dialog box. Change this value if you want to
insert the row(s) above a different row. Click OK. An empty row (or rows) will be
inserted.
If you select more than one entire row before opening the Edit menu and selecting
Insert rows, the number of rows and beginning row number will be displayed in
the Insert Rows dialog box. Clicking OK will insert the specified number of blank
rows above the specified beginning row. Select multiple rows in either of two
ways:
Click and drag the mouse pointer over the row numbers of the rows you want
to insert. OR
Select a single row (by clicking on its row number) and then click another
row number while holding down the <SHIFT> key. That row and all
intervening rows will be selected.
You can also delete one or more rows in the Editor. Select one or more rows, as
described above, then open the Edit menu and select Delete rows. The Delete Rows
dialog box will appear, asking you to confirm the deletion. In the Delete from Row
box, enter the number of the first row to delete. In the To Row box, enter the
number of the last row you want to delete. Or you can select To End of File to
delete all remaining rows. Click OK to delete the row(s), or Cancel to cancel the
deletion.
Tagging Data
Quality Analyst lets you tag numeric data for “special” treatment. You can tag a
single value or an entire row, or you can tag all values of a variable that meet
criteria you specify.
Note: Configuring how Quality Analyst treats tagged data globally is explained
under Missing and Tagged Data on page 271. You can override the
global setting at the variable level as explained under Tagged Data
Handling on page 287.
Quality Analyst uses the values of the key columns specified for your
Data Set to keep track of which data are tagged. So if you intend to tag
data, be sure to specify key columns for your Data Set. See Specifying
Key Columns for the Data Set on page 57 for instructions.
Tagged data is displayed with an asterisk (“*”) next to the data in the cell.
Use the same procedure to untag a tagged value (this function is a toggle). Also,
see Removing Tags on page 85.
2 Open the Edit menu and select Tag Rows. The Tag Rows dialog box will
appear.
To tag all rows below the beginning row, select To End of File. (If you
select this option, new rows that you add to the data file will not be
tagged.)
5 Click OK.
Tagged rows appear gray on your screen, and asterisks appear next to the row
numbers and the numeric data in each row.
You cannot make any changes to data in a tagged row. If you want to do so, you
must untag the row first.
Select the variable you want to tag from the drop-down list labeled Variable. The
description of the variable will appear in the adjacent text box.
Just below the variable, specify the criteria by which you want to tag data.
To tag all values of the selected variable that are above a certain value, enter that
value in the text box labeled Above. This is usually used to tag high outliers. To
include instances of that exact value, select “=” from the drop-down list. To tag
only values greater than that, select “>” from the drop-down list.
To tag all values of the selected variable that are below a certain value, enter that
value in the text box labeled Below. This is usually used to tag low outliers. To
include instances of that exact value, select “=” from the drop-down list. To tag
only values less than that, select “<” from the drop-down list.
If you enter both an Above criterion and a Below criterion, Quality Analyst will tag
values thus:
Normally, Variable is selected in the List by area. This causes the Variable drop-
down list to display variables rather than columns. That is, subgrouped
measurements are listed as a single item. With this setting, any value in any
subgroup of the selected variable that meets your criteria will be tagged. If you
select Column instead, the drop-down list will list individual measurements in
subgrouped variables individually, allowing you to tag individual measurements
separately.
Removing Tags
To remove all tags (on values and rows) in a Data file, open the Edit menu and
select Clear Tags.
2 In the Find box, enter the text you want to search for.
3 In the Search box, select By Rows to search each cell in a row before moving
to the next row. Select Current Column to search the column containing the
selected cell.
4 In the Search Direction area, specify whether to search Up from the current
cell toward the top of the file or Down from the current cell toward the bottom
of the file.
5 In the Matching area, select Case Sensitive to find only those cells where the
text and the case of the text matches your Find text. Select Entire Cell if you
want to find only entire cells, not partial cells, that match your find text. For
example, if you select this option and your find text is “BIN1,” Quality
Analyst will find only cells whose entire content is “BIN1” but not cells
whose entire content is “BIN14” or “PARTBIN1.”
6 Click Find Next to select the next cell that meets the criteria specified in the
dialog box.
2 Specify the Find text and Search parameters as explained in the previous
procedure.
3 In the Replace With box, enter the text that should replace the found text in the
Data Set.
4 Click Find Next to find the Find text in the Data Set.
5 If the Find text is found, do one of the following:
Click Find Next to retain this occurrence of the found text and find the
next occurrence.
Click Replace to replace this occurrence of the found text with the
Replace With text. Repeat until you have made all desired changes in the
Data Set.
Click Replace All to replace all remaining occurrences of the Find text
with the Replace With text.
The existing content of the cell, if any, will be deleted and you will be in edit
mode with no input mask displayed. You are free to enter anything you want
(keeping the variable type in mind).
Note that overriding the input mask is an editing function, not an editing mode—it
lasts for the current cell only. The next time you try to enter or edit a cell with an
input mask, it will be enforced. If you find yourself overriding frequently, you
may want to modify or delete the input mask. Changing an input mask has no
effect on existing data; they will not be checked for compliance with the mask, nor
will they be converted to a form that does comply.
Freezing Columns
If your Data file has many columns—enough that some columns extend beyond
the right edge of the Editor window—the first few columns of data may scroll off
the left side of the window as you enter data in the rightmost columns. If those
leftmost columns contain data that help orient you in the Data file—such as date,
time, shift, etc.—it is easy to lose your place when they disappear from view. For
example, the figure below shows a 12-column file with a cell in the last column
selected. Notice that columns 1–4 are not visible.
Quality Analyst lets you “freeze” any number of leftmost columns so that they can
not scroll off the left edge of the window as you move to the rightmost columns in
your file.
To freeze columns:
1 Open the View menu and select Freeze Columns. The Freeze Columns dialog
box will appear.
2 In the box labeled Last Frozen Column Number, enter the number of columns
you want to freeze.
3 Click OK.
In the Editor, a vertical red line will appear along the right edge of the last frozen
column, indicating that all columns to the left of that line are frozen.
In the figure below, the first two columns are frozen, leaving DATE and TIME
displayed for reference. A vertical red line separates the frozen columns from the
remaining columns. Columns 3–7 are not visible, having scrolled “underneath” the
frozen columns as the cursor was moved into the last column.
Many-column Data file with two frozen columns and a cell in the last column
selected
To “unfreeze” columns, open the View menu and select Freeze Columns, select the
No Columns Frozen option, and click OK. All columns will now scroll as you move
around the data grid.
Autocopy will apply when you have selected a cell in a new, empty row, either at
the bottom of the file or after inserting a row. If the selected cell is in a column of
a variable of type D (DateTime), T (time), A (alphanumeric), or N (numeric text),
and you press <ENTER> or <TAB>, the value from the cell immediately above the
selected cell will be copied into the selected cell before the keystroke selects the
next cell. If the column is any other data type, a missing-data symbol (an asterisk,
by default) will be placed in the cell.
Autocopy is a toggle; to turn it off, open the Edit menu and select Autocopy.
Note: Two items on the Data menu—Query Database and Copy SQL to
Clipboard—are available if you are connected to an external database. For
details, see pages 580 (Query Database) and 572 and 579 (copy SQL).
Filtering Data
As you work with data in the Editor, you may find it convenient to display,
analyze, and chart a filtered subset of your data—that is, only those rows with data
that meet certain criteria. In Quality Analyst, you can do this by opening the Data
menu and selecting Filter. The Data Filter dialog box will appear.
In this dialog box, you can specify up to three sets of filter criteria with which to
filter your data. Each filter is displayed and specified in its own input area labeled
Filter 1, Filter 2, and Filter 3. When you specify filters and click OK, only those
rows that meet all three sets of criteria will remain displayed in the data grid, and
will be used in calculations.
Data filters affect only the display and analysis of data. The data rows excluded by
the filter are not deleted from the file. They are, however, excluded from analysis
and charting.
By default, data rows that do not meet your filter criteria will be displayed in light
gray text. This corresponds to the Gray selection in the area labeled Rows Not
Meeting Filter Criteria. If you want to hide the rows, select Exclusive instead.
Specifying a Filter
To specify a filter, you select the data column whose values will be used to filter
your data, and then specify either a single value or a range of values for that
column.
To select the data column, click the drop-down list labeled Column, which will
contain the name of every column in the file. When you select a column, its
description will appear in the adjacent recessed box.
To specify a single value for your filter criterion, select the Match option. The
adjacent recessed box will become active. In it, enter the value to use as a filter
criterion. For example, if you select the column FOREIGN, enter 3 in the Match
box, and click OK, all rows in which FOREIGN equals 3 will be displayed; all
other rows will be hidden.
To specify a range of values to display, select the From and To option. The
corresponding edit boxes will become active. In the From box, enter the lower
“end point” of the range. In the To box, enter the upper end point of the range. All
rows with values (in this column) between these two values will be displayed
(unless excluded by another criterion).
The drop-down lists next to each input box allow you to specify whether to
include or exclude the end points in the filter criterion. To include the lower end
point, select “=” from the list next to From; to exclude it, select “>.” To include the
upper end point, select “=” from the list next to To; to exclude it, select “<.”
To specify a range of values bounded at one end only, simply leave the edit box
for the other end point blank. For example, if you want to display all rows in
which the values in a column are greater than 2.5, enter “> 2.5” for From, and
leave the To box blank. Similarly, if you want to display all rows in which the
values in a column are less than 2.5, leave the From box blank and enter “< 2.5”
in the To box.
Match Filters
The following rules apply to Match filters based on a DateTime variable:
If you specify a date without a time element as the match criterion, all records
with that date, regardless of time, will be retrieved.
If you specify the date with a time element, only records with that date and
that exact time will be retrieved.
If you omit the seconds element from the time element, Quality Analyst
uses 00.
From/To Filters
The following rules apply to From/To filters based on a DateTime variable:
If you specify a date without a time element as the From = criterion, Quality
Analyst uses a time of 00:00:00.
If you specify a date without a time element as the From > criterion, Quality
Analyst uses a time of 23:59:59.
If you specify a date without a time element as the To = criterion, Quality
Analyst uses a time of 23:59:59.
If you specify a date without a time element as the To < criterion, Quality
Analyst uses a time of 00:00:00.
If you omit the seconds element from the time element, Quality Analyst
uses 00.
Multiple Filters
You can specify up to three filters to apply to your data. Only rows that meet all
three criteria will be displayed. For example, before the filters specified in the
previous figure are applied, DEFECTS.DAT looks like this in the Editor:
DEFECTS.DAT after applying the filter (rows in gray text do not meet filter
criteria)
Notice that the status line at the bottom of the window contains the message Filter
Applied. This message is particularly useful if you chose to hide the filtered rows
as in the next figure. Notice, in this figure, that the original row numbers are
retained; Quality Analyst does not renumber the surviving rows.
Filtered Data Set with filtered rows hidden (using the Exclusive setting)
Removing a Filter
To remove a filter, open the Data menu and select Filter, then click the Clear
button. All three column designations will revert to (None). Click OK and all rows
will be displayed in the Editor data grid.
Specify appropriate rows for Control Limit Recalculation Range on the Data
Ranges tab of the Variable Parameters dialog box (for the variable you will be
charting).
On the Center Line and Limits tab of the Control Chart Parameters dialog box,
create a limit region (for the variable and chart type you want) that
corresponds to the Control Limit Recalculation Range.
Select that region and click the Calculate from Data button. Control limits from
all work shifts will be calculated and entered.
Specify a data filter in which the variable SHIFT matches 2.
Create the chart.
Click Save Only Filtered Data to do just that. Click Save All Data to ignore the filter
and save all data to a file.
Recalculating Data
If your Data Set contains one or more calculated variables, you may want to
recalculate the formulas to make sure your data is current before performing a
charting or analysis function. This should only be necessary if another user or
some other computerized process appends to your Data file while you are working
on it.
To recalculate a Data Set, open the Data menu and select Recalculate.
Sorting Data
Quality Analyst lets you sort the current Data Set based on one or more columns
of alphabetic, numeric, time, or DateTime data. You can, for example, sort data
into historical order by sorting on DATE and TIME (or on DATE, if the values in
that variable include a time element). You can sort the data “in place” (under some
circumstances) or send the sorted data to a new file. To sort the current Data Set,
open the Data menu then select Sort. (If this menu item is disabled, see File Paths
and Permissions on page 269 or contact your system administrator.) The Multi-
Key Sort dialog box will appear.
Select the data columns on which you want to base your sort. (See Selecting
Multiple Variables on page 123.) If you have specified key columns for your Data
Set, those variables will be preselected by default. Be sure to select them in the
proper order: the first data column will be the primary sort key, the second will be
the secondary sort key, and so on.
Click OK. The Multi-Key Sort options dialog box will appear.
Under Output Selection, select Sort current file in place to replace the current data
with the sorted data. (If you have defined key columns for this Data Set, and are
sorting by any other variables, this option is not available.) Select Sort into new file
to keep the current Data file as-is and create a new file with the sorted data. If you
select this option, click Select File to specify a name and location for the new file.
In the area labeled Sort Order, select either Ascending (A–Z, 0–9) or Descending
(Z–A, 9–0).
If your Data Set includes a DateTime variable (type D) that includes a time
element, Quality Analyst will properly sort data within each day based on time. If
the time element is absent from any DateTime value, Quality Analyst assumes it is
00:00:00 (midnight at the beginning of the day).
Verifying Data
As you enter data, Quality Analyst does not make sure that your entries conform
to the requirements of the data types you are entering. So before you save your
data to a file, you should verify that the data is properly formatted.
To verify your data, open the Data menu and select Verify. Quality Analyst will
start at the selected cell and proceed first to the right by column, then down by
row, checking the contents of each cell for conformance to the format
requirements for the applicable data type. If it finds an error, Quality Analyst will
select the offending cell and display a message box that describes the error. Click
OK to remove the message box, then edit the cell contents as needed to fix the
problem.
Note: Verification of data begins with the selected cell and proceeds to the right
and then down by row. If you want to verify all data in your Data Set, be
sure to select the first cell in the Editor before opening the Data menu and
selecting Verify. Press <CTRL+HOME> to select the first cell in the
Editor.
Quality Analyst finds only one error at a time, so you should repeat this process
until Quality Analyst finds no more errors. If you open the Data menu and select
Verify with the first cell in the Data Set selected and Quality Analyst selects the
last cell in the Data Set without displaying the error message, then your entire
Data Set should be free of format errors. Note, however, that although Quality
Analyst is reasonably thorough in identifying invalid data formats, it is not
infallible. You may need to visually inspect your data or charts for anomalies.
In the list labeled Column, select the data column you want to fill with random
numbers.
Normally, Column is selected in the List by area. This causes the Column list to
display columns rather than variables. That is, subgrouped measurements are
listed as individual measurements. With this setting, only the individual
measurement columns you select will be filled with random numbers. If you select
Variable instead, the list will display subgrouped variables as a single item,
allowing you to fill the entire subgroup with random numbers at one time.
In the box labeled Number of Values, enter the number of random numbers you
want Quality Analyst to generate. Each random number will be inserted into a row
for the data column you selected. If the Data file contains more rows than the
number you enter here, the remaining rows will retain their current values. If the
Data file has fewer rows than the number you enter here, Quality Analyst will
create new rows for the additional random numbers and fill the other columns
with the missing data symbol.
In the area labeled Data Type, specify whether you want the random numbers to be
Integers or Real numbers.
In the area labeled Distribution Type, select Uniform to have Quality Analyst
distribute the random numbers uniformly, or evenly, between the Minimum and
Maximum values you enter in the adjacent boxes. Check the Without replacement
box to generate all unique random numbers with no duplications. (This is
available for integers only.) If you uncheck this, the generated random numbers
may not be unique. Select Normal to have Quality Analyst generate the random
numbers on a normal distribution curve. If you select this option, enter the desired
Mean and Standard Deviation for the distribution.
Quality Analyst accepts data via DDE through the following mechanisms:
DDE Link: DDE communications are established between a DDE “Server”
application such as Excel or most “wedges” and a DDE “Client” (Quality
Analyst). Data are placed on the Clipboard by the DDE Server for the DDE
Client to read. Only commands (not data) are sent via DDE. (This mechanism
is also known as a “hot link” or “paste-link.”)
“Direct”: Both commands and data are sent via DDE. The sending
application (the “Server”) and the receiving application (the “Client”) must
agree on the format of the DDE “string” at least to the extent that the DDE
Client can locate and read the data.
4
Choosing names for the three identifiers in the DDE “account” is left to the server application.
In general, Service is the name of the application; Topic is a region or zone within the
application, or an arbitrary topic name; and Item is the name or location of an individual
datum.
If the string of information from the DDE server contains more information than
you need, or if it contains multiple values (to be placed in more than one cell), you
can define a DDE Filter to apply to the incoming string. The data will be extracted
and distributed according to the filter definition. See DDE Filter Definition
below.
In this dialog box, you define Quality Analyst as the DDE client and your data
source as the DDE server. You can also define DDE Filters when necessary.
The box labeled Filters lists all defined filters. When the dialog box first appears,
no filters are defined so the Filters list is empty and most of the rest of the dialog
box is grayed out.
Click New to create a new filter. The number of the filter (1, 2, 3, and so on) will
appear in the Filters list. Each filter includes a Start and a Stop definition, which
tell Quality Analyst where to start and stop reading the desired information.
In the area labeled Start, specify the starting position of the usable value in the
DDE string.
If you select At position, enter the number of the character column to begin
reading the value. This option is useful for “fixed field” strings where the
value always starts at the same position in the string.
Select On occurrence of to tell Quality Analyst to begin reading the
value at a particular occurrence of a particular character or string. For
example, if the value starts after the second fence character, enter “2” and “|”
in the two boxes. If you want Quality Analyst to treat the character or string
as part of the value that follows, check the Included box.
In the area labeled Stop, specify the ending position of the value in the DDE
string.
If you select At End Of Data, Quality Analyst will include the entire remainder
of the DDE string in the data.
Select Take next characters to include a specific number of characters in
the value, and enter that number in the box. This is useful if the character
positions of the data elements of a DDE string remain constant.
Select On occurrence of to tell Quality Analyst to stop reading the
value at a particular occurrence of a particular character or string after the
Start of the value. If you want Quality Analyst to treat the character or string
as part of the value, check the Included box.
The delimiting characters or strings you enter in filter definitions need not be
standard ASCII printing characters. You can enter non-printing characters such as
TAB. To represent non-printing characters, use the following table:
Code Description
\E Escape (ASCII 27)
\N New Line (line feed, ASCII 10)
\T Tab (ASCII 9)
\B Backspace (ASCII 8)
\R Carriage return (ASCII 13)
\F Form feed (ASCII 12)
\\ Backslash
\^ Caret
\### Decimal ASCII character ###, for example \032 for space
\O### Octal ASCII character ###
\X## Hexadecimal ASCII character ##
^X <CTRL+X>, where X is any character between @ and underscore,
inclusive. Case is ignored.
The following two filters would successfully extract the “6.45” from our first
example above. The filter on the left would work generally if this positioning is
consistent. The filter on the right is preferable if the positioning might vary
because it extracts everything between “SD” and “E.”
The following filters would successfully extract the date and two measurements
from the DDE string in the second example above.
Note in Filter 2 that the Start and Stop settings are identical—the first occurrence
of “|.” This does not retrieve an empty string, as you might first think, because the
“counter” for the position of the Stop character starts after the position of the Start
character.
When you send these commands, the case of the string is insignificant but all other
aspects of the command must match these specifications—spelling must be exact
and spaces may not be included.
If Quality Analyst receives a string that it recognizes as a “bad” DDE command—
such as “Mode(”—it ignores the command. If it receives a string that it does not
recognize as a command, it treats the string as data. For example, the string
“Mode” would be treated as data and entered into a text cell as data and into a
numeric cell as missing data.
A DDE Example
This section describes a hypothetical example of data entry using DDE. This
example situation is intended to be common enough that these instructions will be
useful as-is, and explicit enough so that you can extrapolate the ideas to your own
different situation. In this example, we assume:
You have an electronic measurement device, such as a digital scale or caliper,
connected to your PC.
You have software on your PC that can read data from your electronic
measurement device and transmit it, using DDE, to other DDE-enabled
software (such as Quality Analyst). This software is often called “wedge
software,” or just a “wedge,” probably because it is “wedged” between the
measurement device and the destination software, in this case, Quality
Analyst.
Briefly, the procedure for using DDE with Quality Analyst is:
1 Set up DDE transmission in your “wedge” software.
2 Set up DDE reception in Quality Analyst.
3 Transfer data from the instrument to Quality Analyst via DDE using the
“wedge.”
4 Terminate DDE connection.
6 In the area labeled DDE Server Control enter the Service, Topic, and Item
identifiers for the DDE “account” (from step 2) under Link To.
Tip: If you copied the datum to the Clipboard in step 2, the Paste Link button
will be available. Click it and the three identifiers will appear in their
respective boxes.
7 Check the Server Enabled box to allow Quality Analyst to retrieve data from
the server application.
8 Click OK to insert the first datum from the DDE connection.
Entering Data
Once you have established the DDE connection, you can use the measurement or
instrumentation device and the “wedge” to enter data into the Quality Analyst
Editor. Here are some typical steps.
1 In Quality Analyst, select the cell into which you want to read a linked datum.
2 Switch back to the server application (the “wedge”). Depending on the
measurement device and “wedge” software, “read out” new measurements.
To do this you may press a button on a digital caliper, press a button on a
digital scale, press a key on the keyboard, or click a button on the screen with
your mouse.
3 In Quality Analyst, the datum will be inserted into the cell, and the “next” cell
in the data grid will be selected. (What the “next” cell is depends on whether
you are in Row mode or Column mode and whether your Data Set has any
calculated variables. See Moving Around in the Data Grid on page 75.)
4 While entering data via DDE, when the cursor selects a cell that should not
accept a linked datum (like a DATE or TIME field), switch to Quality Analyst
and manually move to the next data-entry cell.
Navigate to the folder where you want to save the Data Set. In the File Name box,
enter a name for the Data Set. You do not need to type in “.DAT” for the file-
name extension; Quality Analyst adds that for you. Click OK to save the Data Set.
Starting with version 6, Quality Analyst Header files are stored in XML
(eXtensible Markup Language) format with the file-name extension .NWH. If you
need to save the Data Set for use on a system that uses an earlier version of
Quality Analyst, open the File menu and select Save As Quality Analyst 5.2 HED
File. The Header file (with the extension .HED) will be saved in the same folder as
the original Data Set.
Printing Data
With Quality Analyst, you can print your data either as a quick and simple data
listing or as a formatted report of selected variables with optional summary
statistics.
In the Rows to Print area of this dialog box, specify the rows of data that you want
to print. Select the File Default option to select the rows specified at the file level
(on the File Parameters dialog box), shown in the adjacent recessed box. Select All
Rows to print all rows in the file. Select Last Rows to specify the number of
rows at the end of the file to print. Enter the number of rows in the box. Select
From Row: To Row: to specify the starting and ending row to print. This
lets you select any contiguous set of rows in the file. In the box labeled Number of
Rows per Page, enter the maximum number of data rows to print on a page.
Quality Analyst will start a new page after this many rows, even if more would fit.
If you request more rows than would normally fit on the page, this setting is
ignored.
In the Columns to Print area of the dialog box, specify the columns of data that you
want to print. Select All Columns to print all columns in the file. Select From
Column: To Column: to specify the starting and ending columns to print.
In the Data Set Information area of the dialog box, specify whether you want to
Print Variable Definitions, Print Column Definitions, or Print File Parameters on the
report.
In the Print Title area of the dialog box, specify the text and alignment of a title to
print at the top of each page of output. Enter the text in the edit box labeled Title.
The default title is the File Title specified under File Parameters. Below the title,
next to the label Alignment, specify whether to Left Justify, Center, or Right Justify
the title on the page.
The check boxes on the right side of the dialog box toggle several formatting
options on or off. Select Show Row Number to print the number of each data row
at the left of each row, as displayed in the Editor. Select Show Column Name to
print a header for each column as displayed in the Editor. Select Gray Rows &
Columns to print the row numbers and columns names, if selected, on a gray
background. Select Show Border to print a thin line around the outside of the
entire table of data. Select Show Grid to print thin lines between each cell in the
Data file.
If you want to print only part of your Data file, you can specify rows and columns
in the Print Data dialog box as just described, or you can select a portion of your
file before invoking the Print Data dialog box. Any time you invoke the Print Data
dialog box with anything more than a single cell selected, the Rows to Print and
Columns to Print settings in the Print Data dialog box will reflect that selection.
To begin printing a formatted statistics report, open the Report menu and select
Formatted Data. The Report dialog box will appear.
will be unavailable and the file will be created using default format settings—a
simple row-by-row print-out of all records with a column for each data column.
Click Select to select a Report Format file other than the default using Windows’
file-selection dialog box. (If you do you can skip the next section.) Click Edit to
change the settings contained in the selected Report Format file. Click Create to
create a new Report Format file. (Creating and editing a Report Format is
explained in the next section.)
In the area labeled Assumed Printer Characteristics, enter the number of
Characters per line that the print function should assume. Since most Windows
fonts are variable pitch, this number is only approximate. You may have to
experiment with this value before finalizing your report format.
In the area labeled Selected Columns and Formats, specify how you want to
display each data column you have included in the report:
Two rows of the report are used to label each data column. In the box labeled
Label Line 1, enter the label to place on the first of the two rows. In the box labeled
Label Line 2, enter the label to place on the second of the two rows. By default, the
first line is blank and the second line contains the variable name.
In the box labeled Column Width, enter the width, in characters, of the column for
this data column. Since most Windows fonts are variable pitch, this number is
only approximate. You may have to experiment with this value before finalizing
your report format.
In the area labeled Numeric Format, check Use Default to display values in this data
column using the default number format for the data type. If you uncheck this box,
enter a format string in the adjacent box. (See Specifying Numeric Formats
below.)
Check Break on data change in this column to start a new report section any time
the value in this data column changes. This is particularly useful for variables like
SHIFT or DATE. In each section of the report, summary statistics ( if selected)
will be computed and displayed for the data in that section.
Any other character included in the format string will be printed “as is” at its point
of inclusion. However, several characters may not be used in format strings
because they are used for other purposes in Quality Analyst. Those characters are:
& (ampersand)
* (asterisk)
\ (backslash)
, (comma)
$ (dollar sign)
! (exclamation point)
+ (plus)
If a number is too large to be represented by the format string, it will be preceded
by a percent sign (%) on the report. For example, 91.35 printed using the format
#.## would appear as %91.35.
Normally, negative numbers are printed with a leading minus sign which occupies
the first character space. Specify a trailing minus sign by placing a “-” at the end
of the format string.
CHAPTER 4
Selecting Variables
One task common to many of Quality Analyst’s procedures is selecting variables.
Whether you are creating charts or performing statistical analyses, Quality Analyst
will frequently ask you to specify a single variable or a set of two or more
variables. Sometimes you will need to specify combinations of single and multiple
variables, depending on what you are asking Quality Analyst to do for you.
This chapter describes how to select variables. The methods described here will
apply to any similar variable-selection dialog boxes in Quality Analyst. When
procedures described in this manual call for you to select variables, they will refer
you to this section.
To select a variable:
Click the variable name in the list on the left, OR
Enter the name of the variable in the edit box above the list on the left, then
press <ENTER>.
The variable name will appear in the box above the list.
If the entire dialog box requests only one variable, you can double-click the
variable name in the list on the left to select the variable and launch the next step
(as if you also clicked OK.)
In some Quality Analyst routines (Sorting Data, for example), the order of the
variables in the Selected list will affect the results of the procedure. To make sure
the variables end up in the proper order, add them to the Selected list in the proper
position. Newly added variables are always inserted into the Selected list just
above the variable (or blank row) that is highlighted in the Selected list.
Some Quality Analyst routines can accept a single variable when multiple
variables are allowed, but some require two or more. If you select a single
variables and the routine requires more, a message box will tell you that at least
two variables are required.
List By
In the box labeled List By, select Variable to display the variables in variable order.
Select Column to display each column in your Data Set.
If any of the variables in the file are in rowwise subgroups, each element of the
subgroup will be displayed individually; the variable name will be followed by a
colon and the measurement number (for example, GAP:2). And if you have
rearranged your columns (as described under Specifying Column Definition and
Calculations on page 61) the columns will be displayed in that rearranged order.
Sort By Name
Check Sort By Name to sort the lists of variable or column names alphabetically.
Leave this unchecked to display the variable or column names in the same order as
they appear in the Data file.
CHAPTER 5
When you create run charts, scatter diagrams, quantile plots, and Pareto diagrams,
you must select the chart type first, and then specify the variable or variables to
chart.
Editor with the file BOARDS.DAT loaded and the variable NUM_BDS selected
Not all variable chart types are represented by a button on the toolbar. The
following table lists the single and paired chart types available in Quality Analyst
and shows the corresponding button and menu selection for each.
List of Variable Chart Types and Corresponding Buttons and Menu Selections
Specify the sample size as explained under Subgroup Size and Type–Variable
Data on page 288. The sample size you enter will be assigned to the variable as if
you had assigned it as a Variable Parameter.
Select a subgroup or individual data point for locating the Cumulative Sum V-
mask.
Select Last Subgroup to place the V-mask on the last subgroup or individual data
point in the file.
Select Subgroup to place the V-mask on the subgroup (or individual data
point) that you specify. In the adjacent box, enter the number of the subgroup or
individual data point for placing the V-mask. For files containing rowwise
subgroups, this will be the row number in the file. For files containing fixed-size
columnwise subgroups, this will be approximately equal to the row number in the
Data file divided by the subgroup size. If your file contains individual
measurements in variable-sized subgroups, you may want to select Last Subgroup,
view the chart once to determine the number of the desired subgroup, then come
back and enter that number here.
Select No V-mask to suppress the V-mask display.
Note: The Selected list on the right displays a default set of statistics which you
can specify, as explained under Capability Report Statistics on page 273.
To select additional statistics and codes, highlight them in the Available list and
click Select, or double-click them in the Available list.
To place a new item in a specific place in the Selected list, first highlight the
existing item in the Selected list above which you want to insert a new item from
the Available list. Then select the new item. It will be inserted into a new row; no
items will be overwritten.
In general, variables will be listed down the report page in rows and statistics will
be listed across the page in columns. The <Page Break> code lets you control how
many columns appear on the report. Any statistic just above a <Page Break> code
in the Selected list will be the last column on that page of the report. Because the
display widths of the statistics and other display parameters vary, you may want to
experiment before finalizing the layout of a report.
Breakdown
In the area labeled Breakdown, select Breakdown by Variable Only to break down
the report (down the page) by variable only. Select Breakdown by to group values
of the selected variable according to the value of another variable, and create a
separate section of the report for each grouping. This can be useful if the second
“breakdown” variable is DATE or SHIFT or MACHINE. If you select this option,
select the breakdown variable from the adjacent drop-down list.
If you select a DateTime variable for the breakdown variable, the area labeled
Date Breakdown Options will become available. Here you can specify whether to
group by Day, by Week, or by Month. If you select by Week, specify (using the
adjacent drop-down list) whether you want the week to start on Sunday or
Monday.
If you select a breakdown variable that is not a date variable, check Sort by
categories to sort the grouped sections by the breakdown variable. If you leave
this box unchecked, the sections will be displayed in the order in which they are
found in the file. (DateTime variables are always sorted chronologically.)
Output Type
Quality Analyst can display process capability reports as text in the Graphics
Viewer and it can create a Data Set that contains the process capability data. You
can choose to create either or both types of output.
In the area labeled Output Selection, select the Text Display check box to display
the capability report as text in the Graphics Viewer.
Select the Data File check box to create a Data Set that contains the process
capability data. Click the adjacent Select File button to specify the name and
location of the Data Set.
The following table lists the attribute chart types available in Quality Analyst and
shows the corresponding button and menu selection for each.
nonconformities c-chart
If you try to create an attribute chart of a variable for which you have not declared
a sample size, the Sample Size dialog box will appear.
Pareto Diagrams
Quality Analyst lets you create both single-variable and multiple-variable Pareto
diagrams, or frequency diagrams.
Multiple-variable Pareto diagrams are appropriate for Data Sets in which each row
is a subgroup or sample and each column contains the number of occurrences of
each type of defect.
Multiple Variables
To create a multiple-variable Pareto diagram, either click the Pareto diagram
button on the toolbar, open the Attribute menu, point to Pareto, then select Multiple
Variables. The Pareto Diagram multiple-variable-selection dialog box will appear.
Single Variable
To create a single-variable Pareto diagram, open the Attribute menu, point to
Pareto, then select Single Variable. The Pareto Diagram single-variable-selection
dialog box will appear.
Select one variable as explained under Selecting a Single Variable on page 121.
Click OK to create the chart.
Note that you can create a Pareto diagram from all types of data, not just defect or
attribute data. If your data is one of the other data types, the bars of the Pareto
diagram will represent the number of occurrences of unique values of the selected
variables. For example, if you select a measurement variable, the Pareto diagram
will include a bar for each unique value of that variable in the analysis range. The
size of each bar will represent the number of times that value of the measurement
was recorded.
Note: See Overview of Assignable Causes and Corrective Actions on page 275
for additional information.
4 In the Include area, select the information to include on the report. (Note that
the Comment option refers to comments you may have entered as you entered
cause or action, not the comments that appear on the chart itself.)
5 In the Format area, select a format for your report. The options are illustrated
below.
6 In the Report AC/CA for area, select Current Data Only to report the causes and
actions only for the data currently in the Data Set. Select Entire Data Set to
report all historical causes and actions, even if the corresponding data are not
all currently in the Data Set.
Note: If you are using a connectivity Data Set, your database query
probably does not retrieve all data into your Data Set. Some causes
and actions in your Data Set may be assigned to unretrieved data in
the database. Selecting Entire Data Set in this step will report those
causes and actions, even though the associated data were not
retrieved. Selecting Current Data Only will report causes and actions
only for retrieved data.
If you are using a standard Data Set, you may have archived old data
and deleted them from the Data Set. Removing old records does not
remove causes and actions associated with those records. Selecting
Entire Data Set in this step will report those causes and actions, even
though the associated data are absent. Selecting Current Data Only
will report causes and actions only for data that remain in the Data
Set.
7 Click OK. The report will appear in the Graphics Viewer in the selected
format.
2 Select the variable whose assignable causes or corrective actions you want to
chart, as explained under Selecting a Single Variable on page 121.
3 Click OK. The Assignable Causes Corrective Action Pareto options dialog box
will appear.
6 In the Report AC/CA for area, select Current Data Only to report the causes and
actions only for the data currently in the Data Set. Select Entire Data Set to
report all historical causes and actions, even if the corresponding data are not
all currently in the Data Set. (See Note on page 150.)
7 Click OK.
4 In the Include area, select the information to include in the Data file. (Note
that the Comment option refers to comments you may have entered as you
entered cause or action, not the comments that appear on the chart itself.)
5 In the Report AC/CA for area, select Current Data Only to include the causes
and actions only for the data currently in the Data Set. Select Entire Data Set
to include all historical causes and actions, even if the corresponding data are
not all currently in the Data Set. (See Note on page 150.)
6 In the Format area, specify the format of the data file.
Select CSV to create a comma-separated-value file.
Select DAT to create a Quality Analyst Data File.
Click the corresponding Select File button to specify the name and
location of the output file.
7 Click OK.
Exception Reporting
Quality Analyst lets you generate “exception reports” which list information about
data that produces one or more of the following types of exceptions:
Control Limit violations
Pattern Rule and Run Rule violations
Specification violations
Each exception report includes a separate section for each combination of variable
and chart type you specify. In each section, each exception is listed and described
separately, including any descriptive information specified in the Data Set
definition (see Description Variables on page 279).
After you have created an Exception Definition file, you can generate reports from
it any number of times; you need not create one every time you want to generate a
report. An Exception Definition file is associated with a single Data Set and
cannot be used with any other Data Set.
2 Click the Create button. The Exception Report Settings dialog box will appear.
Note: If instead of creating a new Exception Definition file from scratch, you
want to use an existing one as a starting point, click Select, then use the Open
dialog box to select the existing Exception Definition file.
3 In the Report Title box, enter a descriptive title for this Exception Definition
file. This title will appear in the File Open dialog box when you later open an
Exception Definition file. It will also appear at the top of printed reports. (If
you leave this blank, the file name will be used instead.)
4 The Save QXP File As box shows the name of the file where the settings you
enter will be saved. By default, this is the name of the Data Set with a .qxp
extension. To change this file name, click the adjacent Select button, then use
the Save As dialog box to select or enter a new file name.
5 In the Exceptions area, select the exceptions—Specifications, Control Limits,
or Pattern Rules—you want to include in the report.
6 In the Report Exceptions for area, specify the time period for the data to be
included in the report.
Select All Data to include all data in the file.
To use only the most recent data, select Last. In the adjacent box, enter
the number of time units to include, and in the adjacent drop-down list
select the time unit—Minutes, Hours, Days, Weeks, or Months. Use the
next two drop-down lists to specify which Date and Time variables to use
when filtering the data.
Note: The Last option is available only if you have specified date and/or
time description variables for this Data Set, as explained under
Description Variables on page 279.
7 The large table in the dialog box lists the combinations of variables and chart
types for which exceptions will be reported.
To add a new combination, click New to display the Exception Report
Chart Properties dialog box.
Select the Chart Type from the drop-down list. Click the Variables button
to select the variables, as explained in CHAPTER 4: Selecting Variables
on page 121. Click OK.
To remove a combination from the list, select it, then click Remove.
To edit an existing combination, select it, then click Properties.
8 If you want to apply more advanced data filtering than the Report Exceptions
for settings allow, click the Define button, then use the Data Filter dialog box
as explained under Filtering Data on page 90.
9 When you are done configuring the Exception Definition file, click OK to
save the definition to the file name you specified. Clicking Cancel will
abandon the settings and not save the file. You will return to the Exception
Report dialog box.
Note: You can also generate exception reports from within Run files. See
EXCEPT on page 437 for details.
view them. If there are more rows of exceptions than will fit in the table, use the
vertical scroll bar on the right of the table to view hidden rows.
2 In the File Name box, enter a name for the file. We recommend not specifying
a file-name extension so that the program can apply an appropriate one based
on the file format. But if you specify an extension, it will be used regardless
of what format you choose.
3 In the Output File Format area, select the format for the file.
Select TXT to produce a text file formatted in the same basic layout as the
on-screen report.
Select DAT to produce a Quality Analyst Data file which is space-
delimited.
Select CSV to produce a comma-separated-value file in which strings are
enclosed in quotes and all strings and values are separated by commas.
Select XML to produce an XML (eXtensible Markup Language) file.
Note: When you select a format, the Files box displays a list of existing files
of that type.
Other Charts
In addition to control charts, you can produce several other types of charts by
opening the Other menu in the Editor. Those chart types are:
Run Chart
Scatter Diagram
Normal Probability Plot
Quantile Plot
Weibull Analysis
Box Plots
List of Other Chart Types and Corresponding Buttons and Menu Selections
Run Chart
A run chart is a simple plot of individual measurements with row number along
the X axis. The control limits and center line are not displayed. The values of the
measurements are scaled on the Y axis. Quality Analyst lets you select up to eight
measurement variables to chart. If you choose a single Y-axis variable, Quality
Analyst will display a single chart with the values of the measurement variable
automatically scaled on the Y axis.
If you choose to display more than one variable, you have two options. You can
display a single chart of all variables scaled to the same Y axis, or you can display
multiple charts with one variable each, scaled individually. In the latter case, the
individual charts will be stacked vertically.
To create a Run Chart, click the Run chart button, or open the Other menu and
select Run Chart. The Run Chart dialog box will appear.
Normally, the X axis of a run chart is simply the number of the data row in your
Data file. This will happen if you check the box labeled Automatic (it is checked
by default). If you would rather scale and label the X axis with the values of one
of the variables in your file, uncheck the Automatic box and select a variable from
the list of X-axis Variables, as explained under Selecting a Single Variable on
page 121.
Select up to eight Y-axis Variables to display in the run chart, as explained under
Selecting Multiple Variables on page 123.
If you selected more than one Y-axis variable, click Options to display the Run
Chart Style dialog box. (This dialog box will also appear if you click OK before
clicking Options.)
Use this dialog box to specify how Quality Analyst should display run charts of
more than one Y-axis variable. Select Single Graph to display all Y-axis variables
in a single graph, scaled to the same Y axis.
Select Multiple Graphs to display separate graphs of each Y-axis variable scaled to
their own Y axis.
Scatter Diagram
Quality Analyst can create a scatter diagram of up to eight Y-axis variables and a
single X-axis variable, scaling both axes automatically.
To create a scatter diagram, click the Scatter button, or open the Other menu and
select Scatter. The Scatter Diagram variable-selection dialog box will appear.
Under X-Axis Variable, select the single variable whose values you want to scale
along the X axis, as explained under Selecting a Single Variable on page 121.
Under Y-axis Variables (1 to 8), select up to eight variable variables to scale along
the Y axis. Click OK to create the scatter diagram.
Quantile Plot
A quantile plot is a special type of scatter diagram used to compare two sets of
data. It is constructed by plotting the quantiles (largest value, second largest value,
third largest value, etc.) of one set of data against the quantiles of the other. The
two sets of data can be variables of subgroup measurements (data type 1,2,3...n),
individual measurements (type I), defect data (type P), sample size data (type Z),
or other numeric data (type O).
To create a quantile plot, open the Other menu and select Quantile. The Quantile
Plot variable-selection dialog box will appear.
Select a single X-Axis Variable and a single Y-Axis Variable as explained under
Selecting a Single Variable on page 121. Click OK the create the chart.
Weibull Analysis
Background
Quality Analyst lets you fit the Weibull or the exponential probability distribution
to data and produce a probability plot. When the distribution is appropriate, the
data on the probability plot will follow the distribution line. You may use
multiply-time-censored (suspended) or failure-censored data.
This Quality Analyst function estimates the two-parameter (shape and scale)
Weibull distribution. For the three-parameter Weibull distribution, you can input
the third parameter (location), but Quality Analyst can not estimate it. The
Weibull distribution can be estimated using either maximum likelihood or median
rank regression. The exponential distribution requires only one parameter (failure
rate).
The Weibull distribution is often used to analyze life-test (reliability) data since
both increasing and decreasing failure rates can be represented. The exponential
distribution is used when the failure rate is constant.
Life test data consists of the number of cycles or period of time that the tested
units survive. When all of the units are tested to failure, the data is called complete
data. If the testing is stopped before all units fail, the data is called censored (or
suspended) data. With censored data, the unfailed units may share a common
survival time (single-censored), or the survival times may vary and be intermixed
with the failure times (multiply-censored).
Weibull Distribution
The Weibull distribution is popular in reliability analysis because it is so flexible;
it can assume a number of different shapes. The parameters of the distribution also
have physical meaning.
The shape parameter (beta) is a number that reveals whether older units are more
reliable or less reliable than younger units. If beta is one, the units have a constant
failure rate; this is the exponential distribution. If beta is less than one, the failure
rate decreases with age; if beta is greater than one, the failure rate increases with
age. When beta is 3.44, the Weibull distribution is a good approximation of the
normal distribution.
The scale parameter determines how far the distribution extends over the time
axis. The scale is in the same units as the data. The scale parameter is the point in
time by which 63.2% of the tested units are expected to fail (if the location
parameter is zero).
The location parameter determines the starting point of the distribution on the time
axis. If the location parameter is positive, units are intrinsically reliable until the
time indicated by the location parameter. If the location parameter is negative, the
units aged or suffered from wear before the test began. Usually, the location
parameter is assumed to be zero, and failures can occur as soon as the test begins.
The location parameter can be estimated by trying different values until the data
on the probability plot follows the distribution line.
Exponential Distribution
The exponential distribution is commonly used because it is so simple. The
distribution assumes a constant failure rate, and only one parameter—the average
failure rate—needs to be estimated. The mean time between failures (MTBF) is
also reported; for the exponential distribution, MTBF is simply the inverse of the
average failure rate.
Data Requirements
The Weibull Analysis requires that the data be in a particular form. The Weibull
analysis can operate on data of types I, O, and GM, and on subgroup data. (See
Data Type on page 37, for a discussion of data types.)
If you have recorded only the lifetimes of failed units, all you need is a single
column of the failure times (complete or singly-censored data).
When only the failure times have been recorded for censored data, Quality
Analyst assumes the unfailed units were all tested for the same length of time
(singly-censored data). When doing a Weibull analysis on this type of data, you
must either specify the lifetime of the unfailed units, or let Quality Analyst use the
lifetime of the longest running failed unit.
If the lifetime of the unfailed units is recorded as well (multiply-censored data), a
second column is used to indicate failed and unfailed units. To mark a unit as
failed, the second column should contain a “1” (one). To mark a unit as unfailed,
the second column should contain a “0” (zero).
You can also use grouped data (type GM). This is useful for large Data Sets. (See
Using Weibull with Large Data Sets below.)
Select the variable that contains the lifetimes of the tested units. Only the names of
appropriate data types will be listed. (Selecting variables is discussed under
Selecting a Single Variable on page 121.) Click OK to display the Weibull Options
dialog box.
Surviving Units
In the area labeled Surviving Units, specify the nature of your data. Select None if
the data you are analyzing represents all failures and no survivors. Select Failure
Identification Variable if your data represents both failures and survivors and you
are using a second variable (containing zeros and ones) to indicate which is which.
Select that variable from the drop-down list. If there are any surviving units that
are not identified individually, select Number of Surviving Units and enter the
number in the adjacent box. If you select this option, also specify the Lifetime of
Surviving Units. Select Use Maximum Failure Lifetime to use the maximum lifetime
of any failure in the Data file.
Distribution
In the area labeled Distribution, select either the Exponential Distribution or the
Weibull Distribution for this analysis. If you select Weibull, also specify the
Estimation Method. Select either Maximum Likelihood or Rank Regression. If you
select Weibull, you must also specify the Location Parameter. Select None to
perform a two-parameter Weibull distribution, or select the edit box and enter a
value for location parameter to perform a three-parameter distribution.
Plot Scale
For both the X Axis and Y Axis, you can impose a scale on the distribution chart.
For each axis, select Default to let Quality Analyst determine the scale, or enter
custom values for Minimum and Maximum.
Confidence Limits
In the area labeled Confidence Limits, specify the confidence interval to use. Select
either None, 95 Percent or enter a custom Percent.
Using grouped data and a separate variable to identify the failed bearing cages,
you can create this Data file.
The first column (LIFE(G)) is the lifetime of both failed and surviving bearing
cages. The second column (LIFE(M)) is the number of units corresponding to
each lifetime. The third column (FAILED) distinguishes between failed units (1)
and surviving units (0). Note that each failed unit with a unique lifetime has a
single entry in the file. The two units that failed after 990 hours have a single entry
with a multiplier of 2. The surviving units are represented by an average lifetime
and the number of units that survived within the range represented by that
average.
In this example, using grouped data lets you represent 1,703 units in a Data file
that has only 9 lines.
Box Plots
A box plot is a plot that gives you a general idea of the distribution of the values
of a variable along a scaled axis. Variables are listed along the Y axis and the
X axis is scaled. In Quality Analyst, the scaled axis can be in measurement units
or can be standardized to the specifications for the variables. The boxes can be
either in quartiles (with the median shown) or standard deviations (with the mean
shown). You can also select statistics to compute and display.
Selecting Variables
To create a box plot, open the Other menu and select Box Plots. The Box Plots
variable-selection dialog box will appear.
Selecting Statistics
In the area labeled Statistics, select the statistics you want to display to the right of
the box plot. Only the variables in the Selected list on the right will be displayed
on the plot.
To select additional statistics, highlight them in the Available list and click Select,
or double-click them in the Available list.
To place a new item in a specific place in the Selected list, first highlight the
existing item in the Selected list above which you want to insert a new item from
the Available list. Then select the new item. It will be inserted into a new row; no
items will be overwritten.
Note: The Selected list on the right displays a default set of statistics which you
can specify, as explained under Box Plot Statistics on page 274.
Breakdown
In the area labeled Breakdown, select Breakdown by Variable Only to display all
values of a variable in a single box plot. Select Breakdown by to group values of
the selected variable according to the value of another variable, and display a
separate box plot for each grouping. This can be useful if the second “breakdown”
variable is DATE or SHIFT or MACHINE. If you select this option, select the
breakdown variable from the adjacent drop-down list.
If you select a date variable for the breakdown variable, the area labeled Date
Breakdown Options will become available. Here you can specify whether to group
by Day, by Week, or by Month. If you select by Week, specify (using the adjacent
drop-down list) whether you want the week to start on Sunday or Monday.
If you select a breakdown variable that is not a DateTime variable, check Sort
Categories to sort the grouped box plots by the breakdown variable. If you leave
this box unchecked, the boxes will be displayed in the order in which they are
found in the file.
Note: You can specify default breakdown parameters for individual columns in
a file. This is explained under Box Plot Parameters on page 332.
Box Type
In the area labeled Box Type, you have three choices. Select Box & Whiskers to
create a plot in which the two middle quartiles of data are displayed as boxes
adjacent to the median value, and “whiskers” extend out to encompass the upper
and lower quartiles. Outliers are displayed beyond the whiskers as asterisks. Select
Mean and +/- 3 Std Dev to create a plot in which three standard deviations of data
above and below the mean are displayed as boxes adjacent to the mean value.
Select Process Capability to create a plot that distributes your data according to the
current process capability parameters. The configuration of those parameters is
described under Process Capability (Histogram) Parameters on page 318.
Units
In the box labeled Units, select Standardized by Specifications to scale and label
the X axis of the box plot with the specification limits for the variables you are
plotting. Select Measurement Units to scale and label the X axis with the actual
values of the measurement units for the variables you are plotting. If you select
this option, check Show Specifications if you want to display the specifications for
the variables as vertical lines on the plot.
Saving Charts
As you create charts, Quality Analyst temporarily saves their Graphics files using
a naming scheme that allows it to keep track of them internally. The Graphics files
will have names like 1000001.NWG or 2.NWG (depending on your license type).
These Graphics files are retained for the duration of the current Quality Analyst
work session, even if you close the plot in the Graphics Viewer.
First, make sure the Drive and the Directory areas at the bottom of the dialog box
display the location to which you want to save the Graphics file.
The upper portion of the dialog box lists, in three recessed boxes, all Graphics
files that have been created during the current Quality Analyst session. The
recessed box in the middle, labeled Graphics Files, lists the temporary names of
the Graphics files (without their extension). The box on the right, labeled Data
File | Variable | Chart, shows which Data file, variable, and chart type are displayed
in the chart. These two boxes are not editable. The box on the left, labeled Save
As, shows the file name that will be used to store the plot when you click the OK
button. “<Don’t save>” indicates that the chart will not be saved—the default
condition.
To assign a permanent file name to each Graphics file for saving, follow these
steps:
Select the plot in the list by clicking its row in any of the three recessed
boxes.
Enter a legal file name (with no extension) in the edit box just below the
Save As label.
Press <ENTER>. (The file name of the chart, with its path and the extension
.NWG added, will appear in the Save As column.)
After viewing the plot, return to the Save Graphics Files dialog box in any of these
ways:
Click any visible portion of the Save Graphics Files dialog box.
Click the Editor button on the toolbar of the Graphics Viewer.
Double-click the Control-menu box in the upper left-hand corner of the
Graphics Viewer window to close it.
Save Graphics Files dialog box with permanent file names assigned
CHAPTER 6
The Graphics Viewer can also be run as a separate program. This section
describes the capabilities of the Graphics Viewer and how to use them.
Quality Analyst uses the Graphics Viewer to display both graphical and some
textual output. Graphical output is any output that is displayed as a chart. This
includes everything on the Variable, Attribute, or Other menus, and the Graphics
Output options of the Regression functions on the Statistics menu. Textual output
is any output that is displayed as a text report in a Graphics Viewer window. This
includes everything on the Statistics menu (but only the Text Output options of the
Regression functions), and (on the Report menu) the Formatted Data report, the
Assignable Cause Corrective Action Report, and the Process Capability report.
When both the Editor and the Graphics Viewer are running you can easily switch
between them:
To switch from the Graphics Viewer to the Editor, open the File menu and
select QA Editor, or click the Editor button (shown at left) on the toolbar.
To switch from the Editor to the Graphics Viewer, open the Run menu and
select Quality Analyst Graphics or click the Graphics Viewer button (shown at
left) on the toolbar.
Tip: If you frequently run the Graphics Viewer by itself, you might want to
create your own desktop icon for the Graphics viewer. Refer to your
Windows documentation for instructions.
The main difference between using the Graphics Viewer by itself and using it
from within the Editor is that you must “manually” open a previously saved
Graphics file. To open a graphics file, open the File menu and select Open, or click
the File Open button on the toolbar. The File Open dialog box will appear.
The Files of type list shows NWG Graphics Files (*.nwg)—the file type that the
Graphics Viewer can open. Select the file you want to display.
Below, the Graphics Viewer is shown displaying HousingXR.nwg which contains
an X-bar and Range chart of the variable GAP in the Data Set HOUSING.DAT.
To view additional charts, simply repeat the File Open process. You can open as
many chart windows as will fit on your screen. That number will depend on the
display resolution of your monitor.
Note: For the sake of simplicity, the rest of this section refers to “chart
windows” even though a chart window can contain either graphical
output or textual output.
arranges all chart windows within the Graphics Viewer window in a way that
emphasizes horizontal orientation. Tiling the windows vertically arranges all
chart windows within the Graphics Viewer window in a way that emphasizes
vertical orientation. With certain numbers of chart windows, therefore, tiling
horizontally and vertically will have the same result.
To cascade chart windows, open the Windows menu and select Cascade or click
the Cascade button.
To tile chart windows horizontally, open the Windows menu and select Tile
Horizontal or click the Tile Horizontal button.
To tile chart windows vertically, open the Windows menu and select Tile Vertical or
click the Tile Vertical button.
You can also arrange chart windows “manually” by dragging the edges or corners
of the restored (that is, neither maximized nor minimized) chart.
The following figure shows the Graphics Viewer window in which two charts
have been minimized and a third and fourth have been tiled vertically.
2 In the Paper area, select the Size of the paper you want to print on and the
Source of the paper. The available options vary from printer to printer.
3 In the Orientation area, chose Portrait (vertical) or Landscape (horizontal).
4 In the Margins (Inches) area, specify the size of the four margins between the
edge of the paper and the chart or report.
5 Click OK.
This is a standard Windows dialog box that lets you specify such settings as the
print range, number of copies, page orientation, and so on. When you are ready to
print the chart, click Print.
Group Layout
To assemble several charts and perhaps a text report on a page for printing, first
assemble them into a Group Layout window in the Graphics Viewer.
This dialog box contains buttons for several alternative combinations of chart
placement and number of charts. The buttons and their names are as follows.
The selected layout appears darkened. Select the layout you want by double-
clicking the corresponding layout button, or clicking the layout button and then
clicking OK. A Group Layout window will be created, and it and all the chart
windows will be tiled within the Graphics Viewer window. The Group Layout
window has an empty panel for each chart you plan to assemble. Two charts, a
text report, and an empty Side-by-Side-Over-Single Group Layout window might
look like this.
1 Enable the Select cursor by clicking the Select toolbar button (shown at left)
or opening the View menu, pointing to Cursor Type, then clicking Select.
2 Click-and-hold the chart area (not the title bar) of any chart window, then
drag the mouse pointer (the appearance of which will change as shown at left)
to any chart panel of the Group Layout window and release the mouse button.
A copy of the chart will appear in the panel, scaled to size.
3 Fill the other panels in the Group Layout window with the other charts using
the same technique.
Dragging and dropping the charts into the Group Layout panels might look like
this.
Once you have placed charts in a Group Layout panel, you can rearrange them by
dragging a chart from one panel and dropping onto another. Notice, though, that
this creates a copy of the dragged chart in the destination panel, replacing the
existing chart, if any, in the destination panel. If you want to replace the
overwritten chart, drag a new copy of it from its individual chart window.
A Group Layout may contain only a single text report. If you include a text report
in a Group Layout, we suggest that you place it in a pane that extends the entire
width of the Group Layout, as shown in the example above. This can help avoid
potential problems that may result from scaling the text to fit the display area.
Notice that when you move the mouse pointer over a border, it changes into a
two-way resizing pointer—a pair of parallel lines with opposed arrowheads. When
you click-and-hold this pointer, you can slide the border in the direction of either
arrow—either up and down or left and right. Release the mouse button to
complete the resizing.
If your Group Layout includes borders that intersect, and you move the mouse
pointer over the intersection of the borders, it will change into a four-way resizing
pointer—two crossing double-headed arrows. When you click-and-hold this
pointer, you can slide the intersection in any direction—up or down or left or
right. Release the mouse button to complete the resizing.
Because the printed page may not have the same proportions as the Group Layout
window when you resize panels, the shapes and sizes of the printed panels may
differ from those of the on-screen panels. If precise sizing is important, you may
have to experiment with different arrangements in the Group Layout window.
Use the buttons at the top of the window to manipulate the preview:
Click Print to print the chart or report from preview mode.
Click Prev or Next to display other pages in a multi-page report.
Click Two Page to view two pages side-by-side. In two-page mode, the same
button will be labeled One Page. Click it to return to the single-page view.
Click Zoom In to enlarge the chart or report within the window. Or you can
click anywhere on the simulated page while the mouse pointer looks like a
magnifying glass.
Click Zoom Out to shrink the chart or report within the window.
Click Close to exit print preview and return to the normal screen display of
the chart or report.
Annotating Charts
Quality Analyst lets you annotate charts by:
modifying or adding chart titles,
attaching text comments to data points on the chart, and
modifying or adding chart footers.
2 Select the title you want to edit using the <UP> or <DOWN> arrow keys.
3 To edit the selected cell, press the <INS> key.
Note: In place of the previous two steps, you can double-click a title.
4 Enter new text or edit the existing text.
5 Press <UP> or <DOWN > to accept the changes and move to another cell in
the list.
6 Repeat steps 2–5 to edit or create additional titles.
7 Click OK.
Note: Quality Analyst uses the values of the key columns specified for your
Data Set to keep track of which data point each Graphics Comment is
assigned to. So if you intend to use Graphics Comments, be sure to
specify key columns for your Data Set. See Specifying Key Columns for
the Data Set on page 57 for instructions.
To label a data point with its row number, the values of its description
variables, or an assigned cause or corrective action, use a callout instead.
See Working with Callouts on page 210 for instructions.
In the edit box, enter the comment you want to add to the chart, then click OK. The
comment will appear on your chart.
You can add comments this way to charts either in individual chart windows or in
panels of a Group Layout window.
Moving Comments
To move a comment that you have added to a chart, select the Comment mouse
pointer. With the I-beam of the Comment mouse pointer, click and drag the
comment to a new location on the chart. Quality Analyst will “attach” the
comment to whichever data point is nearest the leftmost character of the comment
when you release the mouse button.
2 To change any footer text, edit the appropriate cells in the grid.
3 To delete a single cell, select the cell than click Delete. The cell contents will
be deleted, but the contents of other cells will not shift up or left to fill the
space. (If you select a cell and press <DEL>, the contents will be deleted and
you will enter edit mode so you can enter new cell content.)
4 To delete an entire footer line, select the line by clicking its line label (for
example, Line 2), then click Delete. (The remaining lines will not shift up to
fill the space.)
5 When you are done editing footer lines, click OK.
The X and Y values of any point on the chart (not necessarily a data point).
Depending on which details you want to view, you can view them in either of two
ways:
In a single detail window that displays details for whichever data point you
point to with the mouse. (If you print a chart or save it as an image file, the
content of the detail window is not included in the output.)
In one or more callouts that you “attach” to a data point on the chart. (If you
print a chart or save it as an image file, callouts are included in the output.)
Note: You can specify the content of the detail window before or after
displaying the detail window. Any changes you make while the detail
window is displayed will appear instantly.
The figure above illustrates how you can display more description variables in the
detail window than appear along the X axis. The settings that produce such a
result are shown in the File Parameters figure on page 281.
If you choose to display X and Y coordinates in the detail window, you should be
aware that the X value shown will be the row number of the nearest data point, but
the Y value shown will be that of the place in the middle of the mouse cursor,
even if you are pointing to “white space” on the chart. To view the Y value of a
data point, you must center the mouse cursor on that data point. (Even then, the
Y value shown will be approximate.)
You can select one or more of the following pieces of information to appear in
callouts:
the row number of the data points
the values of the description variables you have selected for the Data Set
the cause or corrective action you have assigned to the data point
Callouts are not saved with a Graphics File, but they do appear when you print a
chart or save it to an image file (e.g., JPG or BMP).
Note: To configure the background color for callouts, see Background Color
for Callouts on page 222.
Note: Specifying the content of callouts is not retroactive. If you have added
callouts to your chart, specifying different content will affect subsequent
callouts but not existing callouts.
Note: If you have turned off all callout content (see previous section),
double-clicking a data point has no effect.
If the default location of the callout is not satisfactory, you can move it.
Zooming in on Charts
If a chart contains many, many records, data points may become crowded in the
X-axis direction, making it difficult to interpret the chart or to select individual
points for attaching comments or callouts, or for assigning causes or corrective
actions. In these cases, you may find it helpful to “zoom in” on the data and chart
only a subset of the records in the entire Data Set.
To zoom in on a chart:
1 Click the Zoom In button on the toolbar (shown at left).
2 Click and hold the mouse button down at a place on the chart above the
X-axis where you want the zoomed chart to begin (or end).
3 Drag the mouse right (or left) to the place on the chart above the X-axis
where you want the zoomed chart to end (or begin). As you drag the mouse, a
dotted rectangle will show the portion of the chart that will be included in the
zoomed chart.
4 Release the mouse button. The X-axis of the chart will “stretch” to include
only those records you outlined.
You can also zoom in or out precisely on the X or Y axis by specifying the limits
of either axis.
3 Enter values for X-axis or Y-axis settings as described in the table below.
4 Click OK.
Setting Description
Minimum The minimum data value that can be displayed on the chart.
Maximum The maximum data value that can be displayed on the chart.
Scale Start X axis: The value near or to the left of the left end of the chart at
which to start labeling tick marks.
Y axis: The value near or below the bottom of the chart at which to
start labeling tick marks.
Scale Increment The spacing of labeled tick marks.
Tick Start X axis: The value near or to the left of the left end of the chart at
which to start placing tick marks.
Y axis: The value near or below the bottom of the chart at which to
start placing tick marks.
Tick Increment The spacing of tick marks.
Note: The zoom feature is visual only. None of the calculations of any SPC
parameters, such as control limits, are affected by zooming in.
cause or action to any charted data point of those variables in the Graphics
Viewer.
Note: Quality Analyst uses the values of the key columns specified for your
Data Set to keep track of which data point each cause or action is
assigned to. So if you intend to assign causes or actions, be sure to
specify key columns for your Data Set. See Specifying Key Columns for
the Data Set on page 57 for instructions.
2 To assign a cause to the data point, select a cause from the Assignable Cause
list.
3 To assign an action to the data point, select an action from the Corrective
Action list.
4 If you want to add a clarifying or explanatory comment to the data point,
enter it in the Comment box.
5 Click OK.
The plotting character of the data point will change to indicate that you have
assigned a cause or action to it.
See Specifying Plotting Characters on page 226 for instructions on selecting the
plotting character used for assignable causes and corrective actions.
See Analyzing Assignable Causes and Corrective Actions on page 148 for
instructions on configuring and generating a report of assignable causes and
corrective actions.
Note: The previous two settings are particularly useful if you have a black-only
printer or a monochrome monitor (or any monitor on which colors are
hard to distinguish).
Bar Fill
By default, the bars in bar charts (like histograms and Pareto diagrams) are filled
with a solid color, but you can select a different fill pattern.
You can also specify the colors for each of three “Rainbow Zones”. 5
For each combination there is a square color sample, a Change button, and (except
for rainbow zones) a Line Thickness selector.
5
See Rainbow Zones on page 252 for an explanation.
To specify a color:
1 Click the Change button for the color you want to specify.
2 Select a color using the Windows Color selector.
Specifying Fonts
Quality Analyst lets you specify the fonts used for several types of text on your
charts.
In the Font Category list, select the type of text for which you want to specify a
font. That category will appear in red in the applicable preview panes.
Below the preview panes, select a Font, Font Style, and Size for that font category.
You can see what your font looks like in the Sample area. (Note: The preview
panes do not reflect your font choice; they are intended only to show you where it
will be used on charts.)
When you select X-Axis Label for Font Category, the X-Axis Label Angle list
becomes available. If you want to rotate the labels on the X axis of your charts,
select the angle (in degrees clockwise from horizontal) from the list, or enter a
specific angle in the box using the keyboard.
In the Font list, select the font you want to use for all plotting characters. The grid
below displays the character set for the selected font.
The Control Charts area displays the characters that will be used for each type of
data point on control charts. The Other Graphics area of the dialog box displays
the plotting characters that will be used for data points on other types of charts
available in Quality Analyst (that is, most of the charts found on the Other and
Statistics menus in the Editor). You can specify plotting characters for up to eight
data series.
If you want to suppress the display of plotting characters, so that only the lines
(not the points) of data appear on charts, see Plotting Characters on page 250 for
global settings, or Plotting Characters on page 309 for variable-specific settings
for a Data Set.
2 In the box labeled JPEG Output Quality (1-100), enter an integer between 1 and
100. Higher values produce higher-quality images but larger files. Lower
values produce lower-quality images but smaller files.
3 Click OK.
The printed border will surround all charts, titles, and labels on the printed page. If
you are printing a double chart type (like an X-bar-and-Range chart), a single
border will surround both charts.
Graphics files contain all information about the chart or report needed to
regenerate the chart when you reopen the file with the Graphics Viewer. This
option lets you—or anyone else with Quality Analyst—regenerate the chart and
then manipulate it using the Graphics Viewer (for example, change colors or
fonts, arrange it in a Group Layout, add comments, and so on). 6
Bitmap files contain a static picture of the chart or report as it appears in the
Graphics Viewer when you save it (that is, the colors, fonts, and plotting
characters on your system will be represented). You can open these files with most
graphics- or photo-editing programs, or insert or embed them, for example, in
word-processing or slide-presentation files.
Note: You can save Graphics files from within the Graphics Viewer (as
explained below), but you will probably prefer to do so from within the
Editor, since the Editor’s procedure is more flexible. That procedure is
described under Saving Charts on page 183.
To save a chart or report to a Graphics file from within the Graphics Viewer:
1 Make the window containing the chart or report you want to save active.
2 Click the File Save button on the toolbar, or open the File menu and select
Save As. The standard Windows Save As dialog box will appear. The current
file name will be the temporary file name assigned by Quality Analyst.
6
The settings that control the cosmetic appearance of charts and reports in the Graphics Viewer
(colors, fonts, plotting characters, and so on) are not stored in the Graphics files. Thus, if you
save a Graphics file and give it to other Quality Analyst users, the chart or report will be
displayed with their cosmetic settings, not yours.
You can open a saved Graphics file later with the Graphics Viewer to redisplay
the chart or report. This is explained under Running the Graphics Viewer Alone
on page 188.
To save a chart or report to a bitmap file from within the Graphics Viewer:
Use the procedure above, but in the Save as type list, select one of the bitmap
file formats.
Note: You can specify the image quality (and degree of compression) of saved
JPEG files. See Specifying JPEG Quality on page 228 for instructions.
1 Start the Graphics Viewer, if it isn’t running already, by running the program
NWAGraph.exe. (See Running the Graphics Viewer Alone on page 188.)
2 Start Windows Explorer.
3 Size and position the application windows so that some portion of the
Graphics Viewer is visible on the desktop.
4 In Explorer, display the file list of the Quality Analyst folder (directory), or
any other folder in which you have saved Graphics files.
5 Select any number of Graphics files in the list by clicking the first one, then
clicking additional file names while holding down the <CTRL> key.
6 Click-and-hold the mouse on any selected file, drag the mouse pointer over
any visible portion of the Graphics Viewer, and release the mouse button.
7 Tile or Cascade the windows in the Graphics Viewer. You will see an
individual chart window for each file you dropped onto the icon from
Explorer.
In the target application, pasting the graphic usually involves positioning the
cursor where the graph is to be pasted and then opening the Edit menu and
selecting Paste. Doing this in the WordPad program supplied with Windows
might look like this:
CHAPTER 7
Customizing Charts
With Quality Analyst, you can configure virtually all aspects of the way it
analyzes data and creates control charts. You configure Quality Analyst by
specifying settings, or parameters, at several levels of detail:
System Parameters (page 242) constitute “default” settings and apply unless
they are overridden by file, variable, or control chart parameters.
File Parameters (page 278) apply to all charts created from data in a single
file and remain in force unless overridden by variable or chart parameters.
Variable Parameters (page 283) apply to charts of individual variables
within individual files and remain in force unless overridden by chart
parameters.
Control Chart Parameters (page 295) apply to chart types and sometimes to
specific combinations of variables and chart types within a Data file.
Process Capability Parameters (page 318),
Box Plot Parameters (page 332), and
Pareto Parameters, (page 333)
are detailed chart parameters for specific types of charts.
Short Run Parameters (page 336) apply to the short run form of control
charts of data in a file.
Many of the parameters that you can specify are entered using “tabbed” dialog
boxes similar to this one:
Each category of parameters is presented on its own “tab” that looks much like a
divider in a 3×5-card box. To select a category of parameters, click anywhere on
the tab for that category and it will appear “on top” of the other tabs.
To help you find your way around the configuration and customization parameters
and the tabbed dialog boxes, the following table shows the inputs that are entered
on each. For example, you specify Control Limit Adjustments on the tab labeled
Subgroup/Sample Size and Type on the Variable Parameters dialog box, which you
invoke by opening the Parameters menu and selecting Variable. Pertinent sections
of this table are reproduced at the beginning of the sections on each parameter
type.
System Parameters
This section describes the system-level settings that you can specify in Quality
Analyst. To specify system parameters, click Settings on the main Quality Analyst
start-up screen. This will open the tabbed Quality Analyst Settings dialog box. (If
the Settings button is disabled, see File Paths and Permissions on page 269 or
contact your system administrator.)
The following table shows the input items you can specify on each tab.
Tab Inputs
Tab Inputs
The buttons on the right side of the Settings window are available from all tabs.
Click OK to accept all current settings for the current Quality Analyst session.
If you have changed any settings and later try to quit Quality Analyst, you
will be prompted to save or abandon those changes.
Click Cancel to abandon any changes you have made on any tab since the last
time you clicked Save or OK.
Click Save to save any changes you have made on any tab to the Quality
Analyst configuration file, which is named QA.NWC and is stored (by
default, on Windows XP systems) in the “Documents and Settings\All
Users\Application Data\NWA” folder. QA.NWC is an XML-structured file
which you should not modify except with the Settings program.
Click Save As QA.NWA if you want to save all settings in a configuration file
that is compatible with versions of Quality Analyst earlier than 6.1.
Click Print to print the current settings. The Print Quality Analyst Settings
dialog box will appear. Choose to print the settings on the current tab only, or
on all tabs. Click Printer to select and configure your printer. Click OK to print
the settings.
Click Reset to restore all settings on all tabs to their default, as-installed
values.
General
On the General tab of the Quality Analyst Settings dialog box, you specify several
miscellaneous configuration settings.
Parameter Description
$DATE Current date (not file date).†
Example: 9/23/03
$FILE Drive, path, and name of the file being printed, labeled “File:”
Example: File: C:\PROGRA~1\QA\HOUSING.DAT
$FILEBASE Name of the file being printed, without the drive or path, labeled “File:”
Example: File: HOUSING.DAT
$TIME Current time in 24-hour “military” format.†
Example: 17:35
† The date and time will be displayed according to the settings for Date Format and Time Delimiter
which you can configure elsewhere on the General tab.
Note: You can also use these Replaceable Parameters in File Title, described on page 278.
Date Format
In the area labeled Date Format, you specify the format in which Quality Analyst
prints, displays, and recognizes date data. In the area labeled Style, select the style
for date data.
You can also specify the punctuation symbol used to separate the elements of the
date. Select a period (“.”), slash (“/”), or hyphen (“-”) from the Delimiter list.
In the Two-Digit Years area, enter a two-digit year to specify the 100-year period
that may be represented with two-digit years in Quality Analyst Data Sets. Dates
within the range shown may have two-digit years. Dates outside this range must
have four-digit years or they will be misinterpreted. For example, if you enter 95,
Quality Analyst will interpret the date 9/23/95 as 9/23/1995 and the date
9/23/94 as 9/23/2094. This setting has no effect on dates that have full four-digit
years.
It is important to specify the date format here before importing a Data file that
contains date data. If the date data in the imported file is formatted differently than
specified here, Quality Analyst may not interpret it correctly.
Select Require acknowledgment of Run file Errors if you want Quality Analyst
to pause and request acknowledgment from the user each time a Run file error
occurs.
Select Halt on any Run file Error if you want Quality Analyst to halt the Run
File Interpreter whenever a Run file error is encountered.
Select Continue from nested Run file Errors if you want to cancel the execution
of a nested Run file that produces an error but continue the execution of the
“parent” Run file.
Text Settings
In the area labeled Text Settings, you can specify whether Quality Analyst will
regard the case (upper or lower) of text significant in sorting, extracting, grouping,
and filtering data based on text fields. Check the box labeled Text comparisons are
case sensitive if, for example, you want Quality Analyst to treat “RED” and “Red”
and “red” as different. Clear the check box if you want Quality Analyst to treat
them as identical. By default, this check box is cleared.
For example, if this box is checked, attribute data labeled “Pitted” and “PITTED”
would be grouped into separate bars on a Pareto diagram—probably not your
intent. If this box is cleared, that data would be grouped into the same bar on the
Pareto diagram.
Select the Allow Double-Byte Character Set Text check box if you want Quality
Analyst to allow the use of double-byte characters, as may be required, for
example, to accommodate Asian languages.
Report Date
If your Report Title includes a date code ($DATE), you can override that date by
entering one in the area labeled Report Date.
Decimal Symbol
In the Decimal Symbol area, specify whether Quality Analyst should use a period
or a comma as the decimal point when displaying numbers on the screen and on
printed reports. (In Quality Analyst Data Sets, numbers are always stored with a
period for the decimal point.)
Time Delimiter
Quality Analyst prints, displays, and recognizes time data in a 24-hour “military”
format—hours past noon are numbered 13, 14, 15, and so on. In the area labeled
Time Delimiter, you can specify the punctuation symbol used to separate the hours,
minutes, and seconds. Choose either a period (“.”) or a colon (“:”).
Plotting Characters
In the area labeled Plotting Characters, specify whether Quality Analyst should
display characters for each type of data point on your charts.
Select the check boxes to enable the display of plotting characters for In-control,
Out-of-control, and Rule violation data points. Quality Analyst will always display
data lines, but if you clear any of these check boxes, the plotting character for the
corresponding data points will not appear on the chart lines.
These settings become “global” defaults, but you can override them for specific
variables in your Data Sets, as explained under Plotting Characters on page 309.
Note: To configure the actual plotting characters and colors that Quality
Analyst will use for enabled data points, see Specifying Colors and Line
Thicknesses on page 220 and Specifying Plotting Characters on
page 226.
You have two options for displaying warning or inner limits or both in your
charts:
Select Yes to display them.
Select No to not display them.
In the Show Target on area of the Limits & Specifications tab, you can specify
whether you want Quality Analyst to display the target value for measurement
variables on control charts. To display the target values on each pair of chart types
shown, check the adjacent box.
In the Show Specifications on area of the Limits & Specifications tab, you can
specify whether you want Quality Analyst to display the upper and lower
specifications for measurement variables on control charts. To display the
specifications on each pair of chart types shown, check the adjacent box.
Specification limits and target values for measurement variables are specified
when you define or modify the structure of your Data file as described under
Lower, Upper, and Target Specifications on page 46.
Rainbow Zones
Quality Analyst lets you display and configure “rainbow zones” for your charts—
different colored areas at varying distances from the center line to the limits.
In the area labeled Show Rainbow Zone on, select the corresponding check boxes
to enable the display of rainbow zones on X-bar, I, Median, and EWMA Charts;
Range and Standard Deviation Charts, or P, NP, C, and U Charts. Clear the check
boxes to suppress the display of rainbow zones.
If you enable rainbow zones for X-bar, I, Median, and EWMA Charts; select either of
the two secondary check boxes if you want to Base I and Median Zones on
Specifications or Base X-bar and EWMA Zones on Specifications (instead of on
control limits). (You can override these settings for individual variables in a Data
Set, as explained under Show Rainbow Zones on Charts on page 317.)
In the area labeled Rainbow Chart Zones, you enable the individual rainbow zones
and specify the size of each. You can specify up to three zones. The inner (or
“green”) zone starts at the center line and extends up and down away from it. The
middle (or “yellow”) zone extends up and down from the upper and lower edges
of the inner zone, and the outer (or “red”) zone extends up and down from the
upper and lower edges of the middle zone.
To enable any of the three rainbow zones, select the corresponding check box. In
the edit box for each zone (labeled % distance to Limit), enter the distance from the
starting edge of the zone to the outer edge of that zone as a percent of the distance
to the control limit. 7
Notes: The “green,” “yellow,” and “red” labels for these zones are common
colors for rainbow zones, but you can choose any color for any zone, as
explained under Configuring the Appearance of Control Charts on
page 223.
Also, even if you enable one or more rainbow zones here, you can
suppress their display in the Graphics Viewer, as explained in the same
section.
Chart Parameters
On the Chart Parameters tab of the Quality Analyst Settings dialog box, you can
configure several aspects of the way Quality Analyst analyzes data for producing
control charts.
7
If you choose to base the zones for X bar, I, Median, or EWMA charts on specifications
instead of control limits (in the Show Rainbow Zones on area above), you should interpret
“% distance to Limit” as “% distance to Specification” in this area.
Lower Upper
Control Limit 0.999 0.001
Warning Limit 0.975 0.025
Pattern Rules
Overview
Quality Analyst can apply “pattern rules” to data and highlight violations of those
rules on control charts. Several sets of pattern rules are built into Quality Analyst.
You can use these pattern rules “as is,” modify them to suit your needs, delete
rules you will never need, or even define new pattern rules. You can also tell
Quality Analyst not to apply any pattern rules. Pattern Rules are defined and
specified on the Pattern Rules tab of the Quality Analyst Settings dialog box.
Quality Analyst can apply two types of pattern rules—zone rules and run rules. To
enable the application of pattern rules, check the box labeled Use Pattern Rules.
Zone Rules
Zone rules signal an out-of-control point when N subgroups in a row are inside (or
outside) a zone defined relative to the center line and control limits of the chart.
The size of a single zone is one-third the distance between the center line and the
upper or lower control limit. The boundaries of successive zones lie at multiples of
that distance from the center line. Thus, when 3-sigma control limits are used, the
zone boundaries occur at multiples of one sigma.
Quality Analyst can apply more than one pattern rule at a time. You can define up
to 20 zone rules, but you need not use all of them. The table below lists the 12
zone rules that are predefined in Quality Analyst.
Run Rules
Run rules signal an out-of-control point when a certain number of subgroups in a
row are increasing, decreasing, or alternating. You can define up to ten run rules
in Quality Analyst, but you may apply only three at a time—one each for
increasing, decreasing, or alternating subgroups. If you select more than one of
each type of run rule, only the shortest will be used. For example, if you select
both a “six subgroups increasing” rule and a “seven subgroups increasing” rule,
only the “six subgroups increasing” rule will be applied. The following table lists
the seven run rules that are predefined in Quality Analyst.
6 increasing 8 increasing
6 decreasing 8 decreasing
7 increasing 14 alternating
7 decreasing
Pattern rules are only useful when applied to subgroups that are statistically
independent from on another. Control charts of subgroups that are inherently
correlated—such as moving-average (or range), CUSUM, and EWMA charts—
cannot use the pattern rules.
Caution: Be careful not to use too many rules at once. As the number of rules
increases, so does the number of false out-of-control signals. For example, on an
X-bar chart with upper and lower control limits set at 3.0 sigma, an in-control
process will have about three false out-of-control signals every 1,000 subgroups.
If the Western Electric Rules are also used, the in-control process will have about
eleven false out-of-control signals every 1,000 subgroups. Adding more rules will
further increase the probability of false out-of-control signals.
Or you can double-click the rule in the list. The name of the rule will appear in the
list labeled Zone Rules Used.
You can add several rules to the Zone Rules Used list by selecting several rules
before clicking Select. There are two ways to do this:
To select a set of adjacent rules in a list, click the first rule, then click the last
rule you want while holding down the <SHIFT> key. The first, last, and all
intervening rules will be selected.
To select several nonadjacent rules, click each one in turn while holding
down the <CTRL> key.
After you have selected all the zone rules you want to apply, click Select.
As a convenience, Quality Analyst lets you add all the Western Electric zone rules
at once. These rules are designated with an asterisk (“*”) in the on-screen list of
zone rules and with a bullet (“•”) in the table above. To add all Western Electric
rules to the list of Zone Rules Used, click the * Western Electric button. These rules
will be added to any rules already in the Zone Rules Used list.
To select run rules, copy rules from the Run Rules Available list to the Run Rules
Used list using the same techniques you used to apply zone rules.
In the box labeled Rule Description, enter the text that you want to appear on the
list of Zone Rules Available to identify this rule.
In the area labeled Trigger Zone Rule, specify the set of subgroups to test for
violation of the rule. For example, if you want to trigger a rule violation when
any 3 subgroups in a series of 4 meet the criteria defined in the rest of the dialog
box, enter 3 in the box labeled On: subgroups violating zone and 4 in the box
labeled of: total subgroups considered. If you want to trigger a rule violation
when an entire series of 7 subgroups meet the criteria defined in the rest of the
dialog box, enter 7 in both boxes.
In the area labeled Zone Limits, specify the zone boundary (or boundaries) relative
to the center line. You can specify an upper zone boundary, a lower zone
boundary, or both. To specify an upper zone boundary, check the box labeled
Upper Limit. When this is checked, the adjacent edit box labeled Sigma above
Center Line becomes active. In this box, enter a multiple of sigma to specify the
location of the zone boundary. Specify a lower zone boundary in the same way—
by checking the box labeled Lower Limit and entering a multiple of Sigma below
Center Line in the adjacent edit box.
To place a zone boundary, either upper or lower, above the center line, enter a
value greater than zero for sigma. To place a zone boundary below the center line,
enter a value less than zero for sigma.
Finally, in the area labeled Violation if Subgroups, specify whether violations
should occur inside the zone or outside the zone by selecting Inside Zone or
Outside Zone. If you have specified both an upper and a lower zone boundary,
“inside” and “outside” should be self-explanatory—“inside” means between the
two boundaries. If you specify a single boundary, however, the meanings of
“inside” and outside” are less obvious and are different for upper and lower limits.
The following table lists the effect of Inside Zone and Outside Zone when used
with a single zone limit.
Effect of “Inside Zone” and “Outside Zone” with a Single Zone Limit
To complete the definition of the new zone rule, click the OK button.
Tip: For examples of how to define zone rule patterns, select an existing zone
rule and click the Change button (discussed in the next section) and look
at the settings for that rule.
In the box labeled Rule Description, enter the text that you want to appear on the
list of Run Rules Available to identify this rule.
In the area labeled Trigger Run Rule, specify the number of consecutive subgroups
to test for violation of the rule. For example, if you want to trigger a rule violation
when 7 consecutive subgroups follow a pattern, enter 7 in the box labeled On:
subgroups in a pattern.
In the area labeled Run Rule Pattern, select the pattern you want to define as a
violation:
Select Increasing if you want to flag cases where the plotted value of each
subgroup in the series is greater than the value of the previous subgroup.
Select Decreasing if you want to flag cases where the plotted value of each
subgroup in the series is less than the value of the previous subgroup.
Select Alternating if you want to flag cases where the difference between the
value of each successive subgroup in the series and the preceding subgroup is
alternately greater than zero and less than zero—that is, cases where the plot
has a sawtooth shape.
To complete the definition of the new run rule, click the OK button.
Histogram Statistics
When you create a Process Capability Histogram using Quality Analyst, you can
select statistics to be computed from your data and displayed on “footer” lines
beneath the histogram in three columns—left, middle, and right. On the Histogram
Statistics tab of the Quality Analyst Settings dialog box, you can specify a set of
default statistics.
These will be displayed unless you override them on the Histogram Statistics tab
of the Process Capability dialog box, as described on page 320. When you
installed Quality Analyst, a starting set of default statistics were selected.
Selecting Statistics
The three main areas of this tab allow you to select the statistics you want to
display in each of the three columns of footer lines below the histogram. In each
area, the upper list, labeled Available, lists all available statistics that can be
displayed in that column.
The lower list, labeled Selected, lists the names of the current default statistics for
each column. You specify statistics to display by copying items from the Available
lists into the Selected lists.
Note: Some statistics are based on the specification limits and process target. If
you select those statistics (or specifications) and display a capability
histogram of data for which you have not specified specifications, those
statistics (or specifications) will not appear below the histogram.
Adding Statistics
To add a statistic to a footer line:
1 Select the statistic in one of the lists of Available statistics by clicking its
name. (Use the scroll bar to the right of the list to find the statistic, if
necessary.)
2 Click the Select button beneath that list of statistics.
Or you can double-click the name of the statistic. The name of the statistic will
appear in the corresponding Selected list.
You can add several statistics to the footer line at once by selecting several before
clicking Select. There are two ways to do this:
To select a set of adjacent statistics in a list, click the first statistic, then click
the last statistic while holding down the <SHIFT> key. The first, last, and all
intervening statistics will be selected.
To select multiple statistics from anywhere in one or more of the three lists,
click each one in turn while holding down the <CTRL> key.
If you select statistics from more than one list (left, middle, or right), be sure to
click Select for each list in which you have selected statistics.
Inserting Statistics
Normally, Quality Analyst adds selected statistics to the end of the Selected lists.
You can control the order of the statistics in the list of footer lines by inserting
statistics into a particular position in the list. For example, follow these steps if
you have already selected three statistics to display, and you want to insert a
fourth between the first and second.
Quality Analyst inserts the new statistic above the selected statistic in the Selected
list. To enable you to add a statistic to the end of the list, Quality Analyst places a
blank row below the last statistic in the list. To add a statistic to the end of the list,
select this blank row before adding the statistic.
Removing Statistics
To remove a statistic from the Selected list, select it then click the Remove button
below the list. You can remove several statistics from the list at once by selecting
several, as described above, before clicking the corresponding Remove button(s).
Capability Parameters
On the Capability Parameters tab of the Quality Analyst Settings dialog box, you
specify several parameters affecting the calculation of process capability and the
appearance of the process-capability histogram.
For Ideal Centerpoint, specify whether to base the Cpk, k, and Cpm indices on the
process Target or the midpoint between the specifications (Specification Midpoint).
If the Cpk, k, and Cpm indices are based on the midpoint between the upper and
lower specification limits, the capability indices are penalized for shifting away
from the midpoint. The target is not considered in the calculations. If the indices
are based on the target, the indices are penalized when they stray from the target.
If the target is set at the specification midpoint, the results of these two options are
identical.
For Assumed Distribution, specify whether the indices should be based on the
Normal distribution or the Current (possibly non-normal) distribution. If the
indices are based on the normal distribution, sigma limits are placed equidistant
from the mean (or median) of the data, even if the probability distribution is not
symmetrical. When the indices are based on the current probability distribution,
the sigma factors are set at the same probability level on the current distribution
that the sigma factor would use in the normal distribution.
For example, sigma limits of 3 on the normal distribution occur at the 0.00135 and
0.99865 probability levels. So, to use sigma limits of 3 with a non-normal
distribution, the limits are set at the 0.00135 and 0.99865 probability levels of the
non-normal distribution.
Which combination of these setting you use depends on the situation. Indices
based on the current probability distribution reflect the risk of actually producing
outside the specifications. Indices based on the normal distribution are more
common in process-control literature. Indices using the process target treat the
target as the desired process average and penalize for any deviation from it. Using
the process midpoint is probably more common.
For Distribution Midpoint, specify whether the distribution should be based on the
Mean or Median value of the data.
For Folded/Truncated Distribution, specify whether Quality Analyst should display
the entire Parent distribution or just the Child portion of the distribution.
Sigma Factor
The sigma factor is used to display the upper and lower process-capability sigma
(standard deviation) limits as vertical lines on a process capability chart. A sigma
factor of ±3 is common and makes a good default to enter here. This corresponds
to a natural tolerance of 6 standard deviations.
A sigma factor of 4 is often appropriate when only limited data (less than 30
measurements) or data from a short time period is available. The sigma factor you
enter here is interpreted as a “plus or minus.” Therefore, a sigma of 4 yields a
tolerance of 8 standard deviations.
Confidence Limits
In the area labeled Confidence Limits, specify the confidence limits, in percent, to
use for process-capability calculations. This value will apply unless you override it
on the Process Capability Calculation tab of the Process Capability dialog box, as
explained on page 330.
This setting will apply unless you override it on the Histogram Parameter tab of the
Process Capability dialog box as explained on page 325.
In the area labeled Temporary File Path, specify the path where you want to store
temporary files that are used “behind the scenes” by Quality Analyst. If you leave
this box blank, these files are stored in the same location as your Data files. You
may be able to improve the performance of Quality Analyst by specifying a
location on your network server or workstation that has a quicker response than
the disk where you keep your Data files. (Conflicts may arise if you are using the
single-user version of Quality Analyst and you specify a folder on your network
that is used for the same purpose by another single-user Quality Analyst
installation. Multiple users of the LAN version of Quality Analyst, however, may
safely specify the same network folder.)
In the area labeled Graphics and Text (*.NWG) File Path, specify the path where you
want Quality Analyst to store Graphics files as it creates and displays them. If you
leave this box blank, these files are stored in the same location as your Data files.
Quality Analyst will store temporary files in the location you specify unless you
rename and save them to a different location as described under Saving Charts on
page 183.
To specify a path:
Select the upper check box and leave the path blank to store files in the same
folder as your Data files.
Select the upper check box and enter a path to an existing folder to store files
there.
Select Use System Temporary File Path to use the folder specified by the
environment variable TEMP (in Windows XP) or the TEMP= statement in
your AUTOEXEC.BAT file (in earlier versions of Windows). The name of
that folder is shown in the adjacent gray box.
In the User Permissions area, specify what Quality Analyst should allow users to
do under certain circumstances. The first three settings here specify how Quality
Analyst should behave when a user accesses files (Data files, Header files, and
other configuration files) that are read-only. These three settings have an effect
only when read-only files are encountered.
When Quality Analyst encounters a read-only file while these three check boxes
are clear, it will display a warning message but will let users open and work with
the file. Users can make any allowed changes temporarily (to produce charts and
reports or to perform other tasks), but those changes will not be saved.
When Quality Analyst encounters a read-only file while these three check boxes
are selected, users will not be able to make any such changes.
Select the Prevent Settings Modification check box to disable the Settings
button on the main Quality Analyst start-up screen if QA.NWC (where these
settings are stored) is read-only.
Select the Prevent Data Modification check box to disable any program actions
that would let a user add, delete, or change data if the Data file of the Data Set
is read-only.
Select the Prevent Parameter Modification check box to disable any program
actions that would let a user change any file-specific configuration parameters
such as File Parameters or Variable Parameters if the Header file of the Data
Set is read-only.
Finally, select the Disable File Utilities check box to disable Quality Analyst’s file
utilities (on the File menu). (This setting applies regardless of the read-only status
of any file.)
Note: The User Permissions settings will not take effect until you exit and
restart Quality Analyst.
You can override this global setting for each variable in a Data file, as explained
under Tagged Data Handling on page 287. How to tag data is explained under
Tagging Data on page 81.
the display widths of the statistics and other display parameters vary, you may
want to experiment before finalizing the layout of a report.
External Database
On the External Database tab of the Quality Analyst Settings dialog box, specify
database connectivity settings. Refer to Configuring Connectivity on page 587 for
explanations of these settings.
8
Though you create “matching” lists of causes and actions in each category, we do not force
you to associate causes and actions from the same category with the variables in your Data Set
(as explained under Specifying Lists of Assignable Causes and Corrective Actions on
page 59). This gives you flexibility in configuring lists that meet your specific needs.
If you are working with an existing category, or if you copied an existing category
as a starting point for a new category, the current causes and actions will appear in
the Assignable Causes and Corrective Actions lists. If you are working with a new
category, the lists will be empty.
File Parameters
File parameters apply to all charts created from the current (open) Data file unless
overridden by variable parameters. To specify file parameters, open the
Parameters menu then select File. (If the Parameters menu is disabled, see File
Paths and Permissions on page 269 or contact your system administrator.) The
File Parameters dialog box will appear.
File Title
In the File Title box, enter a description of the content of the file specific enough to
distinguish it from other similar files, if any. This text will be used to label charts
and reports and will appear next to the file name in the File Open dialog box. The
title may be up to 70 characters long, including spaces and punctuation. You can
also include the current file name, date, or time by including any of the
Replaceable Parameters described under Title for Graphics and Reports on
page 246. If you use Replaceable Parameters, be sure to enter them in all
uppercase letters.
File Range
In the area labeled File Range, specify the range of rows in the Data file for
Quality Analyst to analyze and chart.
Select All Rows if you want Quality Analyst to analyze and chart all data
rows.
Select Last Rows to specify the number of rows at the end of the file to
analyze and chart. Enter the number of rows in the box. This is useful if you
have a large file and only need to see the most recent data.
Select From Row: To Row: to specify the starting and ending row to
analyze and chart. This lets you select any contiguous set of rows in the file.
When the To Row box is selected, pressing the <DOWN> arrow will insert
“(Last)” into the box.
This setting becomes a “default” for the file that can be overridden for any or all
variables. This is discussed below under Variable Parameters on page 283.
Description Variables
With Quality Analyst, you can select “description variables” to appear along the
X axis of your chart. At certain intervals along the X axis, tick marks will be
labeled with the values of the variables you select. You can select any
combination of variables to display, including measurement data.
Description variables for particular data points on a control chart can also be
displayed using the detail-view feature of Quality Analyst, which is discussed in
the section Viewing Chart Details on page 206.
In the area labeled Description Variables, select the variables whose values will
label the X axis of control charts and will appear in detail view windows. The list
labeled In File lists all variables in the file. The list labeled Selected lists the
currently selected description variables. When you first display the File Parameters
dialog box while creating a new Data Set, the Selected list will contain the first
variable of data type D (DateTime) and the first variable of data type T (time) in
the Data Set. You can select additional variables in one of two ways:
Double-click the variable name in the In File list, OR
Click the variable name, then click the Select button.
Similarly, you can remove a variable from the selected list in one of two ways:
Double-click the variable name in the Selected list, OR
Click the variable name, then click the Remove button.
Note: All variables in the Selected list will appear in the detail view windows of
the Graphics Viewer when you use the detail view feature of Quality
Analyst (see Viewing Chart Details on page 206.) Only the first few,
however, will appear as labels along the X axis. You can specify how
many will appear as X-axis labels; see Maximum Variables on X Axis
below.
Variables in the Selected list will appear in the same order on the chart as in the
list. You can control this order with the Select and Remove operations. Variables
are added to the Selected list above the currently highlighted variable in that list.
For example, the figure below shows two pictures of the Description Variables
area of the File Parameters dialog box. The left picture shows the variable
MACHINE highlighted in the Selected list and the variable TIME highlighted in
the In File list. The right picture shows what happens if you click the Select
button—the variable TIME is inserted into the Selected list above MACHINE.
Display Interval
The display interval is the frequency at which description variables appear along
the X axis. You have three options for specifying display interval.
If you select All that fit, Quality Analyst will place as many variable descriptions
along the X axis as will fit, given the spacing specified under Characters per
Variable, which is explained above. Other tick marks along the axis will be left
unlabeled.
If you select Description Changes Only, Quality Analyst will place as many
variable labels along the X axis as will fit, but will omit labels that are identical to
the preceding label. Thus, labels will be displayed only when they change.
If you select Every Plotted Points, Quality Analyst will label ticks at the
interval specified. For example, if you select this option and specify “4” plotted
points, Quality Analyst will label every fourth tick mark. Be aware that if the
display width of the number of Characters per Variable is greater than the distance
between the specified number of tick marks, Quality Analyst will skip more tick
marks than you specify here.
Variable Parameters
Variable parameters apply individually to each variable in the file. To specify
charting parameters for variables, open the Parameters menu then select Variable.
(If the Parameters menu is disabled, see File Paths and Permissions on page 269
or contact your system administrator.) The Variable Parameters tabbed dialog box
will appear. The table below shows the input items available on each tab.
Tab Inputs
Data Ranges Analysis and Display Range
Control Limit Recalculation Range
Missing and Tagged Data Missing Data Handling
Tagged Data Handling
Subgroup/Sample Size and Type Subgroup Size and Type–Variable Data
Subgroup Size and Type–Attribute Data
Control Limit Adjustments
Graphics Comments
Description Variables
Other Process Parameters
Each category is presented on its own “tab” that looks much like a divider in a
3×5-card box. To select a category of parameters, click anywhere on the tab for
that category.
Variable
At the top of the Variable Parameters dialog box , above the tabs, is a drop-down
list labeled Variable in which you select the variable in your Data file for which
you want to specify parameters.
The description of the variable is shown in a recessed box to the right. Remember
that you can set these parameters independently for each variable in a file.
To select a variable, click the drop-down-list button, scroll up or down the list if
necessary, and click the variable name.
Data Ranges
Normally, Quality Analyst will analyze and chart all data in a Data file. However,
you can instruct Quality Analyst to analyze and display only some of your data.
To do so, select the Data Ranges tab in the Variable Parameters dialog box.
For each variable in your Data file, you can specify separately the portion of your
data to analyze and display and the portion of your data to use in calculating
control limits.
analyze. In the edit box after To Row, enter the row number of the last row of
data you want to analyze. To specify the last row in the file, press <DOWN> to
insert (Last).
First, a sample size can be constant throughout the file or it can vary from row to
row.
If the sample size is constant, select Sample Size and enter the sample size in
the adjacent box.
If the sample size varies, your file must contain a type “Z” variable whose
value is the sample size for each row. Select Sample Size variable and select
the name of that variable from the drop-down list.
Second, a sample can occupy a single row or can span multiple rows.
If a sample occupies a single row, select (Row) from the drop-down list
labeled Sample based on.
If a sample spans multiple rows, your file must contain a variable (such as
DATE, LOT, SHIFT, and so on) whose values identify each subgroup. A new
sample begins each time the value of that variable changes. Select that
identification variable from the drop-down list labeled Sample Based on.
With multi-row samples, Quality Analyst totals the defects in all rows to
generate a single chart point for each sample.
Note: Samples that span multiple rows may have either constant or varying
samples sizes. If it is constant, the sample size you enter applies to
all of the multiple rows together, not to each row separately.
Tip: If each row in your Data file contains defect data for a single item (a
circuit board or a roll of material, for example), set Sample Size to 1 and
select an identification variable from the Sample Based on list.
Graphics Comments
On the Graphics Comments tab of the Variable Parameters dialog box, you can edit
text comments that you have added to charts of the current variable.
Note that you can only edit these comments (or delete them) while in the Editor.
You can add comments to charts using the Comment button on the toolbar in the
Graphics Viewer, as described under Annotating Charts on page 203.
The Graphics Comments tab displays a two-column data grid with columns
labeled Chart and Comment. The Chart column shows the chart types for which
you have specified graphics comments. The Comment column shows the
comments you have added. To edit a comment, simply select the cell with the
mouse or arrow keys, press <DELETE> to enter edit mode, and edit the contents of
the cell.
To delete a graphics comment from the list, select the row by clicking the row
number on the left, then click the Delete button. A warning dialog box will appear,
asking you to confirm the deletion. Click OK to delete the comment.
Other
On the Other tab in the Variable Parameters dialog box, you can specify several
additional charting parameters for each variable.
Description Variables
On the Description Variables tab, you can override, at the variable level, the
Description Variables (that appear along the X axis of your charts) set at the file
level on the File Parameters dialog box.
Select File Default to retain the file-level defaults, which are displayed in the
recessed boxes below that option. To override the file defaults, select Custom,
then specify all other parameters as described in the sections Description
Variables and X-Axis Description Variables beginning on page 279.
This table shows the tabs and corresponding inputs on the Control Chart
Parameters dialog box.
Tab Inputs
Scale Y-axis Scale
Center Line and Limits Control, Warning, and Inner Limits
Limit Regions
Default Chart Limits Warning limits display by chart type
Inner limits display by chart type
Plotting Characters Mean and Individual Chart variations
Individual Measurements Chart variations
Plotting Characters
Cumulative Sum Type of CUSUM
V-mask Design
Set Process Parameters
CUSUM Protractor
Other Pattern Rules use
Show Rainbow Zones on charts
p-chart Units
u-chart Units
Show Target on charts
Show Specifications on charts
EWMA parameters
At the top of the dialog box, above the tabs, is a drop-down list labeled Variable in
which you select the variable in your Data file for which you want to specify
Chart Parameters.
To select a variable, click the drop-down-list button, scroll up or down the list if
necessary, and click the variable name.
At the top of each tab in this dialog box, just below its label, is a drop-down list
labeled Chart in which you select the type of chart for which you want to specify
parameters for the selected variable.
To select a chart, click the drop-down-list button, scroll up or down the list if
necessary, and click the chart name.
Scale
On the Scale tab in the Control Chart Limits dialog box, you can specify settings
for the scale of the Y axis for charts of each variable.
To specify these parameters, first select the variable and the chart type to which
you want the parameters to apply. The chart types available in the drop-down list
will depend on the data type of the variable you have selected.
Normally, the Y-Axis Scale is set to Recalculate as shown. This directs Quality
Analyst to automatically calculate the scale range and tick spacing from the data to
produce a readable, informative chart. You may, however, want to specify your
own settings for the Y-axis scale, either to create a common scale for directly
comparing sets of charts, or to fine-tune the appearance of the chart.
To specify Y-axis scale settings, first click the Manual option on the Scale tab. The
recessed gray boxes for each setting will become white, indicating that they are
editable. Then select each parameter in turn and type in the values you want. The
following table describes each Y-axis setting and explains generally how Quality
Analyst calculates each.
Before specifying these settings, however, you may want to start with the values
Quality Analyst would have used if allowed to calculate them from the data. To do
this, click the Calculate button at the bottom of the tab. Quality Analyst will insert
the calculated values in their respective cells. Note that all values will be
calculated; any values you might have entered using the keyboard will be
overwritten.
Note: Quality Analyst uses the values of the key columns specified for your
Data Set to keep track of which limits apply to which data. So if you
intend to define limit regions, as explained below, be sure to specify key
columns for your Data Set. See Specifying Key Columns for the Data
Set on page 57 for instructions.
Furthermore, you can specify distinct sets of data rows (called “limit regions”) of
each variable in a Data file for which to display different center lines and limits.
For each variable, you can specify regions and their associated limits separately
for each applicable chart type.
To specify limits or limit regions, first select the combination of variable and the
chart type to which you want the limits and regions to apply. The chart types
available in the drop-down list will depend on the data type of the variable you
have selected.
Select the Range chart type to calculate or enter parameters that apply to Range
charts of measurements in subgroups. Select the Range (with Individuals charts)
chart type to calculate or enter parameters that apply to Range charts of individual
measurements.
Direct Entry
To enter the center line or any limit directly, select the appropriate cell and type
the value you want. Press <ENTER> or <TAB> to accept the entry.
Specify the data rows from which the control limits should be calculated. The
From Row and the To Row in the third option will display the current boundaries of
the limit region. Change these only if you want to calculate control limits from
data outside the region.
In the area labeled Calculation Method, specify whether to use the Average Range
or Average Standard Deviation to calculate the limits. Note: This area of the
dialog box will not appear if the current variable is type XR or XS, or if you have
configured Quality Analyst to calculate control limits using Individual
Measurements, as explained under Calculation Methods for Charts on page 254.
When you are done, click OK. Quality Analyst will calculate the center line and
limits and then insert the calculated values into their respective cells in the column
for the region.
Set to None
To set all limits for a limit region to None, select the region by clicking its column
header, then click the Set to None button. None will appear in each limit cell.
To set individual limits to None, first select the cell. A drop-down-list button will
appear on the right of the cell. Select None from the drop-down list. When the cell
is selected, you can alternatively press <ESC> then the <UP> or <DOWN> arrow
until “None” appears in the cell. To accept the setting, press <ENTER> or <TAB>.
Set to Recalculate
For the last limit region only, you can instruct Quality Analyst to automatically
recalculate the center line and control limits from the data available each time it
charts. To set these limits to Recalculate, select the region by clicking its column
header, then click the Set to Recalculate button.
To set the center line or a limit in the last region to Recalculate individually, first
select the cell. A drop-down-list button will appear on the right of the cell. Select
Recalculate from the drop-down list. When the cell is selected, you can
alternatively press <ESC> then the <UP> or <DOWN> arrow until Recalculate
appears in the cell. To accept the setting, press <ENTER> or <TAB>.
Set to Default
For the last limit region only, the center line and limits can be set (or reset) to their
default values. The default value for the center line and control limits is
“[Recalculate].” The default values for warning and inner limits depend on the
settings on the Default Chart Limits tab, as indicated in the table below. Default
values are displayed in square brackets, for example “[Recalculate]” as opposed to
the “manual” setting “Recalculate.”
To set all limits to their default values, select the region by clicking its column
header, then click the Set to Default button. The default values will be inserted into
each cell.
Default Values for Center Line and Limits in Last Limit Region
To set the center line or a limit to its default values individually, first select the
cell. A drop-down-list button will appear on the right of the cell. Click it to drop
down the list.
Select the value in square brackets (either “[Recalculate]” or “[None]”) from the
drop-down list. When the cell is selected, you can alternatively press <ESC> then
the <UP> or <DOWN> arrow until the bracketed value appears in the cell. To
accept the setting, press <ENTER> or <TAB>.
There is a notable difference between None and [None] and between Recalculate
and [Recalculate]. The bracketed versions are the default values and will change as
those defaults are modified on the Default Chart Limits tab or at the system level.
The unbracketed versions are “hard” settings and will remain in effect until they
are modified on this tab.
Defining Regions
The first time you display the Center Line and Limits tab for a variable, a single
limit region will be displayed. The Beginning Row will be <1 and the Ending Row
will be >(Last), indicating that the region includes every row of that variable. Both
cells will be gray and inaccessible since, with only one region, the entire file must
be used. The remaining parameters will be set to their defaults which vary by
parameter as discussed above.
There are two ways to increase the number of limit regions—adding and inserting.
Note: If you define more than one limit region and expect to occasionally
archive older data, removing it from the Data Set, be sure to identify key
columns for this Data Set, as explained under Specifying Key Columns
for the Data Set on page 57. This lets Quality Analyst link the limit
regions to the actual data and adjust the regions automatically as data is
removed.
It asks you to enter the Beginning Row of New Region. By default, it displays a row
number one greater than the Beginning Row for the current last region. (Accepting
this value would make the current last region one row long and would assign all
remaining rows in the current last region to the new last region.) Enter the
Beginning Row number of the new region you want to create, then click OK.
Before the region is added, the Set New Region Limits dialog box appears,
indicating that the limits for the region about to be created are not set.
In the Set Limits area of the dialog box, specify how Quality Analyst should set
these limits. Select either Calculate from Data, Set to None (Enter limits later), or
Calculate from Process Parameters and click OK.
If you selected Calculate from Data, the Calculate From Data dialog box will appear
before the region is added. Use this dialog box as shown and explained under
Calculate from Data on page 302. After you click OK, a column for the new
region will appear in the Control Chart Limits dialog box. If you define more than
three regions, a scroll bar will appear below the limit region grid.
Inserting a Region
“Inserting” a region lets you create a new region anywhere in your file. When you
click the Insert Region button, the Insert Region dialog box will appear and allow
you to specify the Beginning Row of Inserted Region and the Ending Row of
Inserted Region.
Enter row numbers as needed to define the region you want, then click OK.
As with adding a region, the Set New Region Limits dialog box will be displayed.
Specify how Quality Analyst should set the new limits as just described.
A column for the new region will appear in the Control Chart Limits dialog box. If
you define more than three regions, a scroll bar will appear below the limit region
grid.
If you enter row numbers for the new region that include rows from the end of one
existing region and the beginning of the next, the specified rows will be
reassigned to the new region. The regions created from the remnants of the
original two will retain their respective settings.
If you enter row numbers for a new region that include all rows from any existing
region, that region will be subsumed into the new region and its limit settings will
be lost.
If you enter row numbers for a new region that designate rows from the middle of
a single existing region, the beginning and end of the existing region will become
two separate regions and the rows you specified will become a new region. The
new regions made from the beginning and end of the original region will retain the
settings of the original region.
Deleting a Region
To delete a region, first select the region by clicking the column header, then click
the Delete button. If you delete the last region, its rows will be added to those of
the previous region. If you delete a region other than the last, its rows will be
added to the next region. In both cases, the limits of the “destination” region will
be applied to its new rows.
Resizing a Region
You can adjust the size of regions by directly entering values for the beginning
and ending rows into the appropriate cells. The beginning row of the first region
and the ending row of the last region are fixed at <1 and >(Last) (indicating the
first and last row, respectively) and can not be changed. If you change the value
for the ending row of a region, the beginning row of the next region will
automatically be adjusted. If you change the value of the beginning row of a
region, the ending row of the previous region will be automatically adjusted.
If you enter values that define regions that Quality Analyst can not resolve, a
warning dialog box will be displayed notifying you that one or more regions will
be deleted to resolve the conflict. Click OK to delete the regions, or click Cancel to
return to the Center Line and Limits tab and resolve the conflict manually.
Chart Types
The display of warning and inner limits is specified separately for charts of
process average (X-bar, Individuals, Median, and EWMA charts), charts of
process variation (Range and Standard Deviation charts) and attribute charts (p,
np, c, and u charts). The charts of process average and process variation are
grouped separately because a single chart from one group can be paired with a
single chart from the other to produce a chart pair in which one is displayed above
the other. Quality Analyst lets you specify limits for these individual charts
separately.
The difference between selecting Yes or No and selecting System Default when it
has the corresponding value is that selecting Yes or No will remain in effect even
if you subsequently change the system default. Setting the system defaults is
explained under Limits and Specifications on page 250.
Plotting Characters
Quality Analyst lets you specify plotting characters to be displayed on control
charts to indicate in-control, out-of-control, and rule-violation points. (See
Specifying Plotting Characters on page 226 for instructions.) But the settings on
the Plotting Characters tab of the Control Chart Parameters dialog box let you
specify different plotting characters for certain types of charts for individual
variables.
Because these points are often numerous and therefore often overlap, interpreting
the charts can be confusing. Quality Analyst provides several options for
displaying such points in a way that may make these charts clearer. You can
specify these options separately for Median and Individual charts (in the area
labeled Median and Individual Chart) and any chart pair that includes an Individual
chart (in the Individual Measurements Chart area).
Note The settings in the Median and Individual Chart area apply only to the
individual measurements in M/I charts, not the median values (which use
the standard plotting characters).
Rather than let Quality Analyst plot each measurement in the subgroup with the
same character, you can instruct the program to use the measurement number (that
is 1, 2, 3...n) as the plotting character. To do so, check the Use Measurement
Number as Plotting Character box.
You can also instruct Quality Analyst to spread the individual measurements in
the subgroup out along the X axis slightly. To do this, check the box labeled Offset
Individual Measurement Plotting.
If your Data Set is structured in columnwise subgroups (see page 19) and contains
a suitable variable, you can use the first one to three characters of the value of a
variable to label individual data points. To do so:
Plotting Characters
In the area labeled Plotting Characters, you can specify whether Quality Analyst
should display characters for each type of data point on charts of the selected
variable.
Select System Default to use the settings specified at the system level (as
explained under Plotting Characters on page 250). When this option is
selected, the three check boxes nearby indicate the system-level settings.
Select Custom if you want to override the system-level settings for the
selected variable. Select the check boxes to enable the display of plotting
characters for In-control, Out-of-control, and Rule violation data points. Quality
Analyst will always display data lines, but if you clear any of these check
boxes, the plotting character for the corresponding data points on charts of
this variable will not appear on the chart lines.
In the Plotting Characters area, you can also configure Quality Analyst to use the
first one to three characters of the value of a variable as the plotting “character”
for all charts (not just Individual or Median-and-Individual charts) of the selected
variable. Select First characters of variable, select the number of characters
from the small drop-down list, then select the variable name from the larger drop-
down list. (For example, if you use the first two characters of the variable
OPERATOR, and it contains the values “Tom,” “Dick,” and “Danny,” then the
points will be labeled “To,” “Di,” and “Da.”) Note that the list of variables will
contain all description variables (data types “D,” “T,” “A,” and “N”) in the file—
the only allowable data types for this variable.
Cumulative Sum
On the Cumulative Sum tab of the Control Chart Parameters dialog box, you
specify parameters affecting the creation of cumulative sum control charts.
Type of CUSUM
Quality Analyst can display two types of cumulative control charts. In the Type of
CUSUM area of the dialog box, select Cumulative Deviation (V-mask) to request a
“V-mask” chart of cumulative deviations from target. Select Minus Reference
Value (Numeric) to request a “Numeric” chart of cumulative deviation from target
minus a reference value. Select the Omit SPC Limits and Out-of-Control Symbols
check box to display CUSUM charts without SPC limits or out-of-control
symbols.
CUSUM chart with (left) and without (right) SPC limits and out-of-control
symbols
V-mask Design
In this area of the dialog box, specify the parameters of the cumulative sum
V-mask.
In the V-mask Type area, specify which control limits to use. Select Upper to use
the upper control limit, producing a cumulative sum chart with a one-sided
positive V-mask. Select Lower to use the lower control limit, producing a
cumulative sum chart with a one-sided negative V-mask. Select Both Upper and
Lower to use both control limits, producing a two-sided V-mask.
You can specify a reference value and decision intervals in the V-mask Parameters
area of the dialog box. To Use Custom Parameters, check the box adjacent to that
label. The nearby recessed boxes will become editable, allowing you to enter a
Reference Value (k) and Upper (+h) and Lower (−h) Decision Intervals.
In the area labeled Parameters Expressed In, specify whether the V-mask
Parameters are expressed in Standard Deviation or Measurement Units. Depending
on the settings of other parameters in this tab, Parameters Expressed In may be
selected automatically.
calculate standard deviation, the Calculate Process Standard Deviation dialog box
(shown on page 294) will appear. In this dialog box, specify whether to compute
standard deviation from individual measurements, average subgroup range, or
average subgroup standard deviation.
CUSUM Protractor
In the area labeled CUSUM Protractor, you can select and define a CUSUM
“protractor” for Quality Analyst to display. If you select Use Protractor, Quality
Analyst will display the slopes of several process levels on the left-hand side of
the chart. You must then specify the process levels to display by entering values
for the Starting level, the Stopping level, and Step at which to increment
intermediate levels.
Other
On the Other tab of the Control Chart Parameters dialog box, you specify several
miscellaneous control-chart parameters.
The difference between selecting Yes or No and selecting System Default when it
has the corresponding value is that selecting Yes or No will remain in effect even
if you subsequently change the system default.
You can specify the system default pattern rules as explained under Pattern Rules
on page 256.
p-chart Units
If your Data file contains attribute data, the input area labeled p-chart Units will be
active. Select the units with which to label percent-defective charts:
Fraction Defective
Percent Defective
Parts per Million (ppm)
Parts per Billion (ppb)
u-chart Units
If your Data file contains attribute data, the input area labeled u-chart Units will be
active. Select the units with which to label nonconformities-per-unit charts:
Number
Parts per Million (ppm)
Parts per Billion (ppb)
The following table shows the tabs and corresponding inputs available on the
Process Capability dialog box.
Tab Parameters
Histogram Statistics Statistics to display
Number of Footer Lines
Histogram Parameters Histogram Bars
Calculated Bar Location
Additional Limits
Features Displayed
Probability Distribution
Histogram Scale X-axis Scale
Y-axis Scale
Process Capability Calculation Standard Deviation Calculation Method
Capability Index Calculation Assumptions
Confidence Limits
At the top of the dialog box, above the tabs, is a drop-down list labeled Variable in
which you select the variable in your Data file for which you want to specify
Process Capability Parameters. To select a variable, click the drop-down-list
button, scroll up or down the list if necessary, and click the variable name. Since
process capability histograms apply only to measurement variables, only
measurement variables will appear on the drop-down list in this dialog box.
Histogram Statistics
On the Histogram Statistics tab, you can specify statistics to be computed from
your data and displayed on “footer” lines beneath the Process Capability
Histogram in three columns—left, middle, and right.
Default Statistics
Quality Analyst has a default set of statistics that you can choose to display. These
defaults are set at the system level (as explained in the section Histogram
Statistics starting on page 262).
To use the current default set, select the Use Default Statistics option near the top
of the Histogram Statistics tab. If you select this option, any statistics you may
have already selected (as described below) will be replaced by the defaults.
However, if you now select the Select Statistics option, the default statistics will
remain and you may remove from or add to any of the three lists.
Selecting Statistics
The three main areas of this tab allow you to select the statistics you want to
display in each of the three columns of footer lines below the histogram. In each
area, the upper list, labeled Available, lists all available statistics that can be
displayed in that column. The lower list, labeled Selected, lists the names of the
current default statistics for each column. You specify statistics to display by
copying items from the Available lists into the Selected lists.
Adding Statistics
To add a statistic to a footer line:
1 Select the statistic in one of the lists of Available statistics by clicking its
name. (Use the scroll bar to the right of the list to find the statistic, if
necessary.)
2 Click the Select button beneath that list of statistics.
Or you can double-click the name of the statistic. The name of the statistic will
appear in the corresponding Selected list.
You can add several statistics to the footer line at once by selecting several before
clicking Select. There are two ways to do this:
To select a set of adjacent statistics in a list, click the first statistic, then click
the last statistic while holding down the <SHIFT> key. The first, last, and all
intervening statistics will be selected.
To select multiple statistics from anywhere in one or more of the three lists,
click each one in turn while holding down the <CTRL> key.
If you select statistics from more than one list (left, middle, or right), be sure to
click Select for each list in which you have selected statistics.
Inserting Statistics
Normally, Quality Analyst adds selected statistics to the end of the Selected lists.
You can control the order of the statistics in the list of footer lines by inserting
statistics into a particular position in the list. For example, follow these steps if
you have already selected three statistics to display, and you want to insert a
fourth between the first and second.
In the Selected list, click the second statistic to select it.
In the corresponding (left, middle, or right) list of Available statistics, select
the statistic you want to insert.
Click the Select button for that list.
Quality Analyst inserts the new statistic above the selected statistic in the Selected
list. To enable you to add a statistic to the end of the list, Quality Analyst places a
blank row below the last statistic in the list. To add a statistic to the end of the list,
select this blank row before adding the statistic.
Removing Statistics
To remove a statistic from the Selected list, select it then click the Remove button
below the list. You can remove several statistics from the list at once by selecting
several, as described above, before clicking the corresponding Remove button(s).
Histogram Parameters
On the Histogram Parameters tab of the Process Capability (Histogram) Parameters
dialog box, you specify several parameters affecting the appearance of the Process
Capability Histogram. You can specify these parameters individually for each
measurement variable in your file. Select the variable using the drop-down list at
the top of the dialog box.
Histogram Bars
In the Histogram Bars area of this tab, you specify how Quality Analyst should
construct the histogram bars.
Select Recalculate each run to instruct Quality Analyst to compute the starting
value, width, and number of bars from the data each time a histogram of the
variable is displayed.
To specify parameters for Quality Analyst to use, first select the Set Manually
option. The recessed gray boxes for each setting will become white, indicating that
they are editable. Then select each parameter in turn and type in the values you
want.
In the Start box, enter the data value at which you want the leftmost histogram bar
to begin. In the Bar Width box, enter the range of data values you want represented
by each histogram bar. Finally, enter the Number of Bars to display on the
histogram.
Before specifying these settings, you may want to start with the values Quality
Analyst would have used if allowed to calculate them from the data. To do this,
click the Calculate button below the three settings. Quality Analyst will insert the
calculated values in their respective cells. Note that all three values will be
calculated and inserted. Any values you might have entered using the keyboard
will be overwritten.
Additional Limits
In addition to the specification limits and target, you can specify up to two more
limits to be displayed on the process capability histogram. In the Additional Limits
area of the Histogram Parameters tab is a 2×2 data-entry grid. It contains a row for
each of the two additional limits and two columns labeled Text and Value. In the
Text column, enter a short text label for Quality Analyst to use to identify each
limit on the chart. In the Value column, enter the value of each additional limit.
Quality Analyst will display a vertical line at each limit, labeled near the top of the
chart with the corresponding text.
Features Displayed
In the Features Displayed area of the Histogram Parameters tab, specify the
combination of statistical parameters to display on the process capability chart.
Check the Mean box to display a vertical line at the mean value of the data.
Check the Process Capability box to display the upper and lower process
capability sigma (standard deviation) limits as vertical lines on the chart. If you
select this feature, also specify the sigma factor to use in computing those limits.
Select Default to use the system-level default value which is displayed in the
adjacent recessed box. Or select Custom and specify a value in the adjacent box.
Check Distribution to display the distribution curve computed from the data.
Check Histogram to display the histogram bars constructed from the data.
Probability Distribution
In the Probability Distribution area of the Histogram Parameters tab, specify the
type of distribution to display. Click the drop-down-list button to display a list
options from which to choose.
If you select either Truncated Normal or Folded Normal, you will need to fill out
the rest of this input area. First specify, in the box labeled Truncation/Fold Point,
the point at which to fold or truncate the distribution. Next, select either Data is
Greater Than Point or Data is Less Than Point to indicate whether the distribution is
greater than or less than the truncation or fold point.
Histogram Scale
On the Histogram Scale tab of the Process Capability dialog box, you can specify
settings for the scales of the X and Y axes of the process capability chart.
To specify these parameters, first select the variable to which you want the
parameters to apply.
Normally, the scales are set to Recalculate as shown. This directs Quality Analyst
to automatically calculate the scale range and tick spacing from the data to
produce a readable, informative chart. You may, however, want to specify your
own settings for either scale, either to create a common scale for directly
comparing sets of charts, or to fine-tune the appearance of the chart.
To specify scale settings, first click the Manual option on the Histogram Scale tab.
The recessed gray boxes for each setting will become white, indicating that they
are editable. Then select each parameter in turn and type in the values you want.
The table below describes the scale settings for the X and Y axes.
Before specifying either set of settings, however, you may want to start with the
values Quality Analyst would have used if allowed to calculate them from the
data. To do this, click the Calculate button at the bottom of the tab. Quality
Analyst will insert the calculated values in their respective cells. Note that all
values will be calculated and inserted. Any values you might have entered using
the keyboard will be overwritten.
Setting Description
Minimum The minimum data value that can be displayed on the chart.
(Must be 0 for Y axis.)
Maximum The maximum data value that can be displayed on the chart.
Scale Start X axis: The value near or to the left of the left end of the chart at
which to start labeling tick marks.
Y axis: The value near or below the bottom of the chart at which to
start labeling tick marks.
Scale Increment The spacing of labeled tick marks.
Tick Start X axis: The value near or to the left of the left end of the chart at
which to start placing tick marks.
Y axis: The value near or below the bottom of the chart at which to
start placing tick marks.
Tick Increment The spacing of tick marks.
Select the System Default option to use system-level defaults which are shown
below the label for each assumption on the left half of this dialog area. Select the
Custom option to specify each assumption individually.
For Ideal Centerpoint, specify whether to base the Cpk, k, and Cpm indices on the
process Target or the midpoint between the specifications (Specification Midpoint).
For Assumed Distribution, specify whether the indices should be based on the
Normal distribution or the Current (possibly non-normal) distribution. (See
Probability Distribution on page 326.)
For Distribution Midpoint, specify whether the distribution should be based on the
Mean or Median value of the data.
For Folded/Truncated Distribution, specify whether Quality Analyst should display
the entire Parent distribution or just the Child portion of the distribution.
Confidence Limits
In the area labeled Confidence Limits, specify the confidence limits to use for
process capability calculations. Select System Default to use the system-level
setting which is shown in the adjacent recessed box. Otherwise, select Custom and
type the desired value in the adjacent box.
Default Variables
In the upper part of the dialog box, select the variables that you want to be
“preselected” each time you start to create a process capability report from this
file. When you actually create the report, you can add variables or remove
variables, but these will show up automatically in the Selected list when you open
the Variable menu and select Capability Report. Selecting variables is explained
under Selecting Multiple Variables on page 123.
Breakdown Defaults
In the area labeled Default Breakdown, specify the default breakdown settings you
want to be “preset” each time you start to create a capability report. When you
actually create a report, you can change any of these settings. Select Breakdown by
Variable Only to break down the report (down the page) by variable only. Select
Default Variables
In the upper part of the dialog box, select the variables that you want to be
“preselected” each time you start to create a box plot from this file. When you
actually create the box plot you can add variables or remove variables, but these
will show up automatically in the Selected list when you open the Other menu and
select Box Plots. Selecting variables is explained under Selecting Multiple
Variables on page 123.
Breakdown Defaults
In the area labeled Default Breakdown, specify the default breakdown settings you
want to be “preset” each time you start to create a box plot. When you actually
create a box plot, you can change any of these settings. Select Breakdown by
Variable Only to display all values of a variable in a single box plot. Select
Breakdown by to group values of the selected variable according to the value of
another variable, and display a separate box plot for each grouping. This can be
useful if the second “breakdown” variable is DATE or SHIFT or MACHINE. If
you select this option, select the breakdown variable from the adjacent drop-down
list.
If you select a DateTime variable for the breakdown variable, the area labeled
Date Breakdown Options will become available. Here you can specify whether to
group by Day, by Week, or by Month. If you select by Week, specify (using the
adjacent drop-down list) whether you want the week to start on Sunday or
Monday.
If you select a breakdown variable that is not a DateTime variable, check Sort by
categories to sort the grouped box plots by the breakdown variable. If you leave
this box unchecked, the boxes will be displayed in the order in which they are
found in the file.
Pareto Parameters
Quality Analyst lets you specify several parameters for Pareto diagrams of
attribute data or measurement categories. To specify these parameters, open the
Parameters menu then select Pareto. (If the Parameters menu is disabled, see File
Paths and Permissions on page 269 or contact your system administrator.) The
Pareto Parameters dialog box will appear. These parameters apply to the entire
file.
Pareto Format
In the area labeled Pareto Format, you can specify several parameters that affect
the appearance of Pareto diagrams.
You have three options in choosing the Bars to Display. If you select All Bars, your
Pareto diagram will contain a bar for every attribute variable or type found. If you
select Largest Bars, you can specify, in the box, the number of bars to display
(the largest n bars will be displayed). If you select Top Percent, and enter a
percentage in the box, the Pareto diagram will contain a number of bars that will
represent at least that percentage of total defects. For either of the latter two
options, you can choose to display the remaining data in a single bar labeled
“Other” by checking the box labeled Remaining bars as ‘Other’. If you clear this
box, the remaining data will not be displayed.
You can instruct Quality Analyst to Sort Bars by either the Size of Bar or the
Names of the variables. If you select Names here and enter Labels under Multiple
Variable Pareto, bars will be sorted by those Labels.
If you select Cumulative Frequency, Quality Analyst will superimpose a line
showing cumulative frequency over the frequency bars on the Pareto diagram.
If you select Bar Percents, Quality Analyst will label each bar with its percentage
value.
In the column labeled Cost, you can enter a cost for each variable to be used for
weighting the diagram. To instruct Quality Analyst to use these costs to weight the
Pareto diagram, check the box labeled Weight Pareto by Cost just below the grid.
The magnitude of each bar will be the product of the number of defects and the
cost you enter.
Select the variable you want to chart from the drop-down list labeled Short Run
Variable. Select the variable to be used as the Identification Variable from the drop-
down list. Click Read/Sort to generate a list of unique values of the Identification
Variable, sort them, and insert them in the Identification column in the grid.
Note: When you add new values (items) to the identification variable in your
file, you can add them to this list manually, or click Read/Sort to
regenerate the entire list.
In the Nominal column, enter the nominal or target values associated with each
identification code. If necessary, enter variation factors for each in the Variation
column to adjust for differences in variation between items.
Repeat these steps for each variable in the Data Set, then click OK. This stores the
short run parameters with the Data Set (in the Header file).
While you are entering Short Run parameters, you may find the buttons on the
right side of the dialog box helpful.
To insert a new row in the data form, first select the row above which you want to
insert a new row (by clicking the row number), then click Insert.
To delete a row, first select it by clicking its row number, then click Delete.
Click Row Mode or Column Mode to switch editing modes as described under
Moving Around in the Data Grid on page 75.
Click Print to print the Short Run parameters.
CHAPTER 8
Overview
Quality Analyst lets you perform several types of statistical analyses on your data:
One-way ANOVA
Completely Randomized
Randomized Block Design
Chi-square
Bartlett’s Chi-square Statistic
Difference Among Proportions
Equal Expected Frequencies
Unequal Expected Frequencies
Correlation
Kruskal-Wallis Statistic
Descriptive Statistics
t-Statistics
Unpaired
Paired
Single Sample
Pearson Product Moment Correlation
Wald-Wolfowitz Runs Test
Regression
Single-Variable
Multiple Regression
Polynomial Regression
Note: If a data filter is in effect, only those data rows that meet your filter
criteria will be included in statistical calculations. See Filtering Data on
page 90.
General Procedure
All statistical analyses are located on the Statistics menu in the Editor. The general
procedure for performing statistical analysis with Quality Analyst is:
1 Open the Data Set that you want to perform a statistical analysis on.
2 Open the Statistics menu.
3 Select the type of analysis you want to perform.
4 Select the variables to analyze.
5 Click OK to perform the analysis (or specify additional optional parameters).
One-way ANOVA
Quality Analyst can perform analysis of variance on two types of experiment
designs:
one-way designs with equal or unequal subclass numbers (uneven Data Set)
one-way randomized block design: a cross-classification of data by blocks (or
subjects) and treatments. This is also referred to as “repeated measures” or
“subjects-by-treatments” design.
Completely Randomized
To create the Analysis of Variance table for a one-way design with equal or
unequal subclass numbers (uneven Data Set):
Open the Statistics menu, point to One-Way ANOVA, then select Completely
Randomized.
This ANOVA requires a balanced design, where some data is present for each
treatment in each block. However, the data may be “unbalanced” with different
numbers of subjects for each treatment or missing data for some subjects. If a
single data point is missing, it will be estimated, minimizing the error sum-of-
squares. If more than one element is missing, they will be estimated using an
iterative procedure. The degrees of freedom will be modified to account for the
missing points.
Chi-square
Quality Analyst can calculate four types of Chi-square statistics, each of which is
appropriate for a particular type of data:
In the upper half of the variable-selection dialog box, labeled Variance Variable,
select the variable to use for the sample variances as explained under Selecting a
Single Variable on page 121. In the lower half, labeled Degrees of Freedom
Variable, select the variable to use for the degrees of freedom.
The output of this routine consists of Bartlett’s Chi-square statistic and the
significance level.
In the upper half of the variable-selection dialog box, labeled Positive Responses
Variable, select the variable to use for the positive responses as explained under
Selecting a Single Variable on page 121. In the lower half, labeled Sample Size
Variable, select the variable to use for the sample size. This must be a type “Z”
variable.
The output for this routine consists of the Chi-square statistic, the common alpha,
and the significance level.
The output for this routine consists of the Chi-square statistic, the expected
frequency, and the significance level.
The output of this routine consists of the Chi-square statistic and the significance
level.
Correlation
Quality Analyst can calculate the linear correlation coefficients for two or more
variables. This routine produces a correlation matrix for all possible combinations
of the selected variables.
To begin the analysis, open the Statistics menu and select Correlation. In the
variable-selection dialog box, select at least two variables as explained under
Selecting Multiple Variables on page 123.
Click Options to display the Correlation options dialog box. (This dialog box will
also appear if you click OK before clicking Options.)
In the area labeled Treat Missing Data, specify how to treat missing data. Select
Casewise if the Data Set contains no missing data. Data rows with any missing
data will be ignored. If you select Pairwise, all possible pairs of data are used, even
if the row contains some missing data. Correlation coefficients will be calculated
using different numbers of data pairs. Selecting Pairwise will slow down
processing.
In the area labeled Output Selection, select the output destination. Select Text
Display to display the results of the analysis on the screen in a window. (Once
displayed on the screen, the results can be saved to a Graphics file using the
Graphics Viewer.) Select Data File if you want to save the correlation matrix in a
Quality Analyst Data file. You can accept the default output file (OUTPUT.DAT)
or specify another by clicking Select File.
The output of this routine consists of the mean and standard deviation of each
Data Set and the number of data pairs used, correlation coefficient, and
significance level for each possible pair of variables.
If you chose to save the correlation matrix in a Quality Analyst Data file, you can
view it by opening it with the Editor.
Kruskal-Wallis Statistic
The Kruskal-Wallis test is a nonparametric analysis of variance for populations
where the assumptions required by one-way ANOVA are not met.
To begin the analysis, open the Statistics menu and select Kruskal-Wallis Statistic.
In the variable-selection dialog box, select at least two variables as explained
under Selecting Multiple Variables on page 123.
The output for this routine consists of the number of data elements, the Kruskal-
Wallis statistic, and the significance level.
Descriptive Statistics
Quality Analyst lets you compute sample or population descriptive statistics for
one or more variables in a Data Set. This routine computes the following statistics:
95% Confidence Interval for Mean
Arithmetic mean
Coefficient of Kurtosis
Coefficient of Skewness
Coefficient of Variation (%) (1)
Fourth Moment about the Mean
Geometric Mean (2)
Harmonic Mean (2)
Maximum
Minimum
Number of Data Elements (N)
Quadratic Mean (RMS)
Standard Deviation (1)
Standard Error (1)
Sum
Third Moment about the Mean
Variance (1)
To compute descriptive statistics, open a file with the Editor then open the
Statistics menu and select Descriptive Statistics. The Descriptive Statistics dialog
box will appear. Select one or more variables as explained under Selecting
Multiple Variables on page 123.
Click Options to display the Descriptive Statistics Options dialog box. (This dialog
box will also appear if you click OK before clicking Options.)
Use this dialog box to specify additional parameters for computing descriptive
statistics. In the area labeled Calculate Statistics By, select Sample Method to use
“N-1” to calculate the variance, standard deviation, standard error, coefficient of
variation, and third and fourth moments about the mean. Select Population Method
to use “N” in those calculations
In the area labeled Output Selection, select Text Display to display the results of the
analysis on the screen in a window. (Once displayed on the screen, the results can
be saved to a Graphics file using the Graphics Viewer.) Select Data File if you
want to save the statistics for each selected variable in a Quality Analyst Data file.
You can accept the default output file (OUTPUT.DAT) or specify another by
clicking Select File.
The Text Display output includes the title of the file, the description of the
variable(s), and the statistics listed at the beginning of this section.
The Data File output includes a row for each variable you selected and a “column”
for each statistic in the following order, left to right:
Variable name
Number of elements
Sum
Minimum
Maximum
Arithmetic mean
Standard deviation
Variance
Standard Error
Coefficient of variation
t-Statistics
Quality Analyst can calculate four types of t-statistics:
Ordinary unpaired sample t-statistic
Paired sample t-statistic
Single sample t-statistic
Pearson Product Moment Correlation
Use this dialog box to select two variables (representing sample sets from normal
populations) for which you want to compute the common t-statistic. Select two
single variables as explained under Selecting a Single Variable on page 121.
Variable 1 is the first sample set; Variable 2 is the second.
Click Options to display the t-Statistics, Unpaired Options dialog box. (This dialog
box will also appear if you click OK before clicking Options.) Use this dialog box
to specify whether the two sample sets (variables) have the same variance. Also
enter the hypothesized mean difference between the two sample sets in this dialog
box. The default is zero.
Paired
To calculate the t-statistic for a set of paired observations:
Open the Statistics menu, select t-Statistics, then select Paired. The
t-Statistics, Paired dialog box will appear.
Use this dialog box to select two variables (representing paired observations) for
which you want to compute the t-statistic. Select two single variables as explained
under Selecting a Single Variable on page 121. Variable 1 is the first sample set;
Variable 2 is the second.
Click Options to display the t-Statistics, Paired Options dialog box. (This dialog
box will also appear if you click OK before clicking Options.) Use this dialog box
to specify the hypothesized mean difference between the two sample sets. The
default is zero.
Single Sample
To calculate the t-statistic for a Data Set with a known population mean:
Open the Statistics menu, select t-Statistics, and select Single Sample. The
t-Statistics, Single Sample dialog box will appear.
Use this dialog box to select the variable with a known population mean for which
you want to compute the t-statistic. Select a single variable as explained under
Selecting a Single Variable on page 121.
Click Options to display the t-Statistics, Single Sample Options dialog box. (This
dialog box will also appear if you click OK before clicking Options.)
Specify the hypothesized population mean of the variable. If you know the
population variance, select Variance and enter it. Otherwise, select Unknown
Variance.
Note: When the population variance is known, the degrees of freedom cannot
be calculated directly and you should obtain the value from a normal
table.
Use this dialog box to select the two paired variables. Select two single variables
as explained under Selecting a Single Variable on page 121.
Use this dialog box to select the variable you want to test for randomness. Select a
single variable as explained under Selecting a Single Variable on page 121.
Click Options to display the Wald-Wolfowitz Runs Test Conditions dialog box.
(This dialog box will also appear if you click OK before clicking Options.) Use
this dialog box to specify two criteria—A and B—for dividing your data. The
divisions should be mutually exclusive. Not all of the data need be included within
the criteria (the center line may be left out, for example).
For each condition, A and B, select a qualifier from the drop down list (choices
shown below), and enter the corresponding value.
< less than
> greater than
<= less than or equal to
>= greater than or equal to
Regression Analysis
Quality Analyst can perform three basic types of regression analysis:
Single-variable regression (several types)
Multiple linear regression
Polynomial regression
Single-Variable Regression
Quality Analyst can perform four types of regression analysis on a single
independent (X) variable.
coefficients.
Selecting Variables
To perform a single-variable regression analysis, open the Data Set that contains
the variables you want to analyze. Open the Statistics menu, select Regression,
then select Single Regression. The Single Regression variable-selection dialog box
will appear.
Select a single variable (or column) each for the Independent (X) Variable and the
Dependent (Y) Variable. (See Selecting a Single Variable on page 121.)
Click OK. The Single Regression options dialog box will appear. Specify
additional parameters for the regression analysis.
Regression Type
In the area labeled Regression Selection, select from one to four types of single-
variable regression you want to perform. Because of the mathematics involved, 9
some regression types can not be performed on certain types of data. Some of
these restriction are listed in the table below. If Quality Analyst encounters such a
problem, the offending regression type will be ignored. Corresponding graphics
output will be suppressed and regression statistics will be displayed as “---” in text
output.
9
Quality Analyst uses exponentiation to convert some functions to linear form.
Graphics Output
In the area labeled Graphics Output, select Regression Plot to display a graph of
the data points, the regressed curve(s), and the confidence limits (if selected).
Select Residuals Plot to display a plot of the residuals of the regression analysis.
Quality Analyst displays this plot in its own window, separate from the regression
plot, so you can plot residuals without plotting the regression, if you want. The
residuals plot is standardized. That is, the values plotted are actually the standard
deviations of each residual value from the residual mean.
Text Output
In the area labeled Text Output, select Regression Statistics to display the
regression statistics for each regression type selected. If you select this option, you
have the additional option, by checking Include Detail Data, of displaying the
actual and calculated Y values. The output will contain a row for each row in your
Data file, followed by the regression statistics.
Select Residuals File to create a Quality Analyst Data file out of the residuals of
the regression analysis. Specify a name and location for the file by clicking Select
File.
Confidence Limits
In the area labeled Confidence Limits, specify the confidence limits you want
displayed on the Regression Plot.
Select None to suppress the display of confidence limits.
Select Default to use the default settings which are displayed in the adjacent boxes.
Select Custom to override the default settings for this plot only. If you select this
option, enter the desired Probability Level in the text box and specify whether the
Limit Type is based on the Individual Points or the regressed Equation.
Select Shelf Life / Stability to analyze the stability of the product. In the Limits area,
select Upper or Lower to indicate which specification limit you want to investigate.
If you have defined upper or lower specifications for the variable being analyzed,
those values will be used as defaults. You can enter a different value in the
adjacent box. Select the Probability Level check box if you want Quality Analyst to
report the probability level intersection with the specification. Enter the desired
probability level in the adjacent box. Select the Regression line check box if you
want to include the intersection of the regression line and the specification in the
report.
Note: If you selected more than one regression type, no confidence limits will
be displayed.
Multiple Regression
Selecting Variables
To perform a multiple linear regression, open the Data Set that contains the
variables you want to analyze. Open the Statistics menu, point to Regression, then
select Multiple Regression. The Multiple Regression variable-selection dialog box
will appear.
Select a single variable (or column) for the Dependent (Y) Variable. (See Selecting
a Single Variable on page 121.) Select one or more variables (or columns) for the
Independent (X) Variables. (See Selecting Multiple Variables on page 123.)
Click Options to display the Multiple Regression options dialog box. (This dialog
box will also appear if you click OK before clicking Options.) Specify additional
parameters for the multiple regression analysis.
Graphics Output
In the area labeled Graphics Output, select Regression Plot to display a graph of
the data points, the regressed curve, and the confidence limits (if selected).
If you select Regression Plot, you may also select Residuals Plot to display a plot
of the residuals of the regression analysis. (Because Quality Analyst displays the
regression plot and the residuals plot in the same graphics window, the residuals
plot is not available unless you select the regression plot.) The residuals plot is
standardized. That is, the values plotted are actually the standard deviations of
each residual value from the residual mean.
Text Output
In the area labeled Text Output, select Regression Statistics to display the multiple-
regression statistics.
Select Residuals File to create a Quality Analyst Data file out of the residuals of
the multiple regression analysis. Specify a name and location for the file by
clicking Select File.
Confidence Limits
In the area labeled Confidence Limits, specify the confidence limits you want
displayed on the multiple-regression plot. Select None to suppress the display of
confidence limits. Select Default to use the default settings which are displayed in
the adjacent boxes. Select Custom to override the default settings for this plot
only. If you select this option, enter the desired Probability Level in the text box
and specify whether the Limit Type is based on the Individual Points or the
regressed Equation.
Polynomial Regression
Quality Analyst can fit your data to a polynomial equation using a least-squares
regression analysis. The form of the polynomial is:
y = B0 + B1x + B2x2 + B3x3 + … + Bnxn
where y is the independent variable, x is the independent variable, B1 through Bn
are the regression coefficients, and n is the order of the polynomial.
Selecting Variables
To perform a polynomial regression, open the Data Set that contains the variables
you want to analyze. Open the Statistics menu, select Regression, then select
Polynomial Regression. The Polynomial Regression variable-selection dialog box
will appear.
Select a single variable (or column) each for the Independent (X) Variable and the
Dependent (Y) Variable. (See Selecting a Single Variable on page 121.)
Click Options to display the Polynomial Regression options dialog box. (This
dialog box will also appear if you click OK before clicking Options.) Specify
additional parameters for the polynomial regression analysis.
Order of Polynomial
Enter the order of the polynomial in the edit box. This is the “n” in the sample
equation above. This value must be less than the number of data pairs (rows) in
your Data Set.
Graphics Output
In the area labeled Graphics Output, select Regression Plot to display a graph of
the data points, the regressed curve, and the confidence limits (if selected).
Select Residuals Plot to display a plot of the residuals of the regression analysis.
Quality Analyst displays this plot in its own window, separate from the regression
plot, so you can plot residuals without plotting the regression, if you want. The
residuals plot is standardized. That is, the values plotted are actually the standard
deviations of each residual value from the residual mean.
Text Output
In the area labeled Text Output, select Regression Statistics to display the
polynomial regression statistics.
Select Residuals File to create a Quality Analyst Data file out of the residuals of
the multiple regression analysis. Specify a name and location for the file by
clicking Select File.
Confidence Limits
In the area labeled Confidence Limits, specify the confidence limits you want
displayed on the polynomial regression plot. Select None to suppress the display
of confidence limits. Select Default to use the default settings which are displayed
in the adjacent boxes. Select Custom to override the default settings for this plot
only. If you select this option, enter the desired Probability Level in the text box
and specify whether the Limit Type is based on the Individual Points or the
regressed Equation.
CHAPTER 9
Automating Procedures
with Run Files
Most Quality Analyst functions can be automated with “Run files”—scripts
created by you that instruct Quality Analyst to automatically perform one or more
of its functions. Quality Analyst Run files can be used to automate repetitive tasks
and to enable less experienced users to perform complex sequences of tasks. Run
files can be especially useful if you want to:
Perform standard sequences of tasks where a large number of charts are
routinely created from the same data sources.
Create time-consuming printer or plotter graphics output unattended.
Automate the creation of multiple chart layouts.
Set up an automated charting session for workers less familiar with the use of
computers and Quality Analyst.
Create a control chart automatically from another software package by setting
up a Run file to perform the required operations.
Use Quality Analyst as the SQC charting component of an integrated
manufacturing-information system.
Run files are easy-to-create text files, and can be as simple as a series of
commands to print control charts. The largest section of this chapter—Run File
Commands and Parameters beginning on page 386—discusses each Run file
command in detail and provides examples of the required command syntax.
You can execute Run files from within Quality Analyst (selecting the desired Run
file from a list), from icons on the desktop, or from other applications. This is
explained in the section Executing Run Files beginning on page 496.
More advanced Run files can prompt the user for information, and perform
complex work flows including page layout, data filters, and calculations. The
section Creating Complete Task Sequences, beginning on page 490, explains the
creation of complex work flows and includes examples of complete Run files. Run
files have allowed many Quality Analyst users to create SPC/SQC charts as part of
Quality Analyst includes a Run file Wizard that guides you, step-by-step, through
the process of creating a Run file. A friendly user-interface lets you select and
insert commands, parameters, file names, and variable names with point-and-click
ease. To launch the Run File Wizard, open the Run menu and select Automation
(Run File) Wizard.
Note: Run file commands are not case-sensitive. As long as they are spelled
correctly they will be interpreted correctly by the Run File Interpreter. To
enhance readability, some commands are shown in this manual using
mixed case.
Use the following table to find the appropriate command for the function you want
to perform.
Command or
Function Type Function
Parameter
Command or
Function Type Function
Parameter
np-chart ATTRIBUT
Pareto PARETO
p-chart ATTRIBUT
Process Capability Histogram PCAP
Process Capability Report (display) MCAPREP
Process Capability Report (prepare) MPCAP
Quantile plot NORMPLOT
Range XRS, IR
Run chart RUNCHART
Scatter diagram SCATTER
Standard Deviation XRS
u-chart ATTRIBUT
Weibull analysis WEIBULL
X-bar XRS
X-bar & Range XRS
X-bar & Standard Deviation XRS
General Statistics Analysis of variance ANOVA, ANOVAR
Bartlett’s test CHITEST
Box Plots (display) BOXPLOT
Box Plots (prepare) MPCAP
Chi-square statistics CHITEST
Descriptive statistics SUMSTAT
Formatted reporting REPORT
Kruskal-Wallis statistic KRUSKAL
Linear correlation CORVAR
Pearson Product Moment Correlation TTESTS
Regression (multivariable) MULTREG
Regression (polynomial) POLREG
Regression (single-variable) REGRESS
t-statistics TTESTS
Command or
Function Type Function
Parameter
Command or
Function Type Function
Parameter
When executing the Quality Analyst Run File Interpreter from a Windows
command line or program icon, you can include any of several parameters in the
command line to control the way the Run file behaves. These are explained in the
section Command Line Options on page 502.
Note: In the examples in the following sections, some Run file command lines
are too long to fit on a single line in this manual so they wrap to an
indented second line. In your actual Run file, however, you must enter
each command line on a single line. For example, two sample Run file
command lines in this manual might look like this:
The indented second line is actually the end of the first line. The text
editor you use to create your Run files will be capable of creating lines
much longer (usually up to 255 characters) than we can display on a
single line in this manual.
Note: Notable exceptions to this general form are the “Ask” commands
(starting on page 399) which prompt the user for input.
XRS is the name of the Run file command. HOUSING.DAT is the name of the input
file. GAP is the name of the variable. X and R are command parameters
(instructions) that tell XRS to produce both an X-bar and a Range chart. G tells
Quality Analyst to create a Graphics file containing the chart information.
This command line alone would produce no “visible” output. To display the chart
on the screen, you would follow it with a CHART command. To print it, you would
follow it with a PRINT command. A Run file that produces a process capability
histogram in addition to the X-bar and Range chart might look like:
XRS HOUSING.DAT GAP X R G
CHART
PCAP HOUSING.DAT GAP G
CHART
However, because the file name contains a space, the Run File Interpreter will try
to interpret “Ring” as the file name and “Weight.dat” as the variable and the
remaining parameters in the statement will be off by one position. This will
produce an error message.
To avoid this, place the Data file name in quotes:
XRS "Ring Weight.dat" WT X R G
In this case, the Run file command line will execute correctly because Quality
Analyst will interpret everything within the quotes as the file name.
This issue is particularly important if you use Replaceable Parameters to assign
file names in Run files. See Using Replaceable Parameters for File and Path
Names on page 487 for more on this issue.
Quality Analyst Run files can produce three types of output. The available
output options vary with each command and are discussed in the section on each
command.
If a Graphics file is created by a Run file command (by including the “G”
parameter), and the Graphics file name is not specified, the default name
GRAPHICS.NWG will be used.
Also note that the “G” and “W” options create Graphics files but do not display or
print them. This is accomplished by an additional command line containing a
“CHART” or a “PRINT” command or both.
ACCA
The ACCA command produces reports on the assignable causes (ACs), corrective
actions (CAs), and associated comments linked to data points in a Data Set. The
syntax of the ACCA command takes one of three forms depending on the desired
format of the output—text report, Pareto diagram, or Data file.
See Analyzing Assignable Causes and Corrective Actions on page 148 for
detailed explanations of some of these parameters.
Text Report
This form of the ACCA command generates a text report that itemizes assignable
causes, corrective actions, or comments. The syntax for this form of the ACCA
command in a Run file is:
ACCA inpfile format vars info scope output outfile
Example: To generate a tabular report of the ACs and CAs (but not comments)
for the variables THICK and ANGLE in the file FLANGE.DAT, including only
data currently in the Data Set, and save that report to the file FlangeACCA.nwg,
the command line is:
ACCA FLANGE.DAT Tabular THICK ANGLE A C Current W
"FlangeACCA.nwg"
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
Pareto Chart
This form of the ACCA command generates a Pareto diagram of assignable causes
or corrective actions to show which are most common. The syntax for this form of
the ACCA command in a Run file:
ACCA inpfile PARETO var bars other info scope output
[outfile]
Example: To generate a Pareto diagram of the assignable causes for the variable
ANGLE in the file FLANGE.DAT, showing the three largest bars individually
and the remaining bars grouped as “Other,” including all ACs in the entire Data
Set, and saving the diagram to the file FlangeACPareto.nwg, the command line is:
ACCA FLANGE.DAT PARETO ANGLE L 3 R A All G
"FlangeACPareto.nwg"
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
Data File
This form of the ACCA command writes assignable causes, corrective actions, or
comments to a file. The syntax for this form of the ACCA command in a Run file
is:
ACCA inpfile FILE vars info scope output outfile
ANOVA
The ANOVA command performs an analysis of variance for one-way
experimental designs with equal or unequal subclass numbers (uneven Data Set).
The syntax of the ANOVA command line in a Run file is:
ANOVA inpfile vars output outfile
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
ANOVAR
The ANOVAR command performs an analysis of variance for one-way
randomized block designs: a cross-classification of data by blocks (or subjects)
and treatments. This is also referred to as “repeated measures” or “subjects-by-
treatments” design. The syntax of the ANOVAR command line in a Run file is:
ANOVAR inpfile vars output outfile
Example: The command line to analyze the variance of the variables LINE1,
LINE2, LINE3, and LINE4 in the file BAGGER.DAT is:
ANOVAR BAGGER.DAT LINE1 LINE2 LINE3 LINE4 W DEFAULT.NWG
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
APPEND
The APPEND command displays an abbreviated version of the Quality Analyst
Editor and lets you add data to a Data Set. The syntax of the APPEND command
is:
APPEND filename numrecords
The window for the Append editor will display five data rows and is not resizable.
If numrecords is greater than five, a vertical scroll bar provides access to the
additional rows.
Click Continue to append the data records you entered, or click Quit to proceed
using the original Data Set.
Example: The following Run file statement displays the editor as shown below.
APPEND DEFECTS.DAT 7
Notice that the vertical scroll bar indicates the presence of additional rows past the
fifth.
AskAlpha
AskDate
AskInteger
AskNumber
AskTime
Note: There are several Run file commands that prompt the Run file user for
one or more inputs of a specific data type and then assign those inputs to
Replaceable Parameters (see page 484). Run file authors can specify the
on-screen text used to prompt the user for input, or they can use default
text. The user’s input is checked to make sure that it complies with the
requirements of the data type.
Several variants of the Ask commands prompt the user for a lower and
upper limit for a range of values, and several other variants prompt the
user for inputs that are to be used as arguments for the $FILTER
parameter (see page 439) of a charting command.
The basic Ask commands (“Ask” followed by a data type) display a dialog box
that prompts you for an input of a specific data type and assigns that input to a
Replaceable Parameter after making sure that it is a valid input. The syntax of the
basic Ask command is:
command param "prompt" "default"
default the value to display in the input box when the dialog box
first appears
Note: To prompt the user for input to be used with the $FILTER parameter of a
charting Run file command, use the MatchFilter or SelectFilter forms of
the Ask command instead.
Use double quotation marks as shown in the syntax statement. Both prompt and
default are optional, but if you omit prompt, you must also omit default.
You can use empty quotation marks for prompt, in which case standard text will
be used. To suppress the prompt, use a quoted space. The default value, if
specified, will be assigned to the Replaceable Parameter when the user clicks
Continue unless it is first deleted or replaced.
If you omit the optional quoted parameters, this command displays the following
dialog box (varying slightly for the different data types):
Example: The following Run file statement displays the dialog box shown
below.
AskDate ^MatchDate "Enter the date to investigate:"
"10/1/04"
In this example, Quality Analyst will make sure that the input meets the
requirements for the DateTime data type.
AskAlphaMatchFilter
AskDateMatchFilter
AskIntegerMatchFilter
AskNumberMatchFilter
AskTimeMatchFilter
Note: See note at AskAlpha AskDate AskInteger AskNumber AskTime on
page 399.
Use double quotation marks as shown in the syntax statement. Both prompt and
default are optional, but if you omit prompt, you must also omit default.
You can use empty quotation marks for prompt, in which case standard text will
be used. To suppress the prompt, use a quoted space. The default value, if
specified, will be assigned to the Replaceable Parameter when the user clicks
Continue unless it is first deleted or replaced.
If you omit the optional quoted parameters, this command displays the following
dialog box (varying slightly for the different data types):
Read more about Replaceable Parameters starting on page 484 and about
Filtering Datas starting on page 90.
Example: The first of the following two Run file statements displays the dialog
box shown below.
AskAlphaMatchFilter ^MatchInsp "Enter the Inspector’s
name:" "STEVE"
ATTRIBUT BOARDS.DAT PART_MIS N G MISSING.NWG $FILTER =
"INSPECT ^MatchInsp"
If the user accepts the default value “STEVE,” then the string “M STEVE” is
assigned to the Replaceable Parameter ^MatchInsp and the second Run file
command is executed as:
ATTRIBUT BOARDS.DAT PART_MIS N G MISSING.NWG $FILTER =
"INSPECT M STEVE"
AskAlphaRange
AskDateRange
AskIntegerRange
AskNumberRange
AskTimeRange
Note: See note at AskAlpha AskDate AskInteger AskNumber AskTime on
page 399.
The Range forms of the Ask command display a dialog box that prompts you for
two (beginning and ending) inputs of a specific data type and assigns those inputs
to two Replaceable Parameters after making sure that they are valid inputs. The
syntax of the Range form of the Ask command is:
command param1 param2 "mainprompt" "prompt1" "prompt2"
"default1" "default2"
default2 the value to display in the second input box when the dialog
box first appears
Note: To prompt the user for input to be used with the $FILTER parameter of a
charting Run file command, use the RangeFilter or SelectFilter forms of
the Ask command instead.
Use double quotation marks as shown in the syntax statement. All quoted
elements are optional from right to left. That is, you may omit a quoted element
only if you also omit all elements after it. You may, however, use empty quotation
marks for any of the quoted elements. In that case, standard text will be used for
prompts, and defaults will be left blank. To suppress any prompt, use a
quoted space. The default values, if specified, will be assigned to the
Replaceable Parameters when the user clicks Continue unless they are first deleted
or replaced.
If you omit the optional quoted parameters, this command displays the following
dialog box (varying slightly for the different data types):
Quality Analyst will make sure that the inputs meet the requirements for the
DateTime data type and that the first date is before (or the same as) the second.
AskAlphaRangeFilter
AskDateRangeFilter
AskIntegerRangeFilter
AskNumberRangeFilter
AskTimeRangeFilter
Note: See note at AskAlpha AskDate AskInteger AskNumber AskTime on
page 399.
Use double quotation marks as shown in the syntax statement. All quoted
elements are optional from right to left. That is, you may omit a quoted element
only if you also omit all elements after it. You may, however, use empty quotation
marks for any of the quoted elements. In that case, standard text will be used for
prompts, and defaults will be left blank. To suppress any prompt, use a
quoted space. The default values, if specified, will be included in the string
assigned to the Replaceable Parameters when the user clicks Continue unless they
are first deleted or replaced.
If you omit the optional quoted parameters, this command displays the following
dialog box (varying slightly for the different data types):
Read more about Replaceable Parameters starting on page 484 and about
Filtering Data starting on page 90.
Example: The first of the following two Run file statements displays the dialog
box shown below.
AskDateRangeFilter ^DateFilter "Enter the date range to
investigate:" "Start of investigation period:"
"End of investigation period:" "10/1/2004"
"10/31/2004"
XRS FLUID.DAT VISC X S G $FILTER="DATE ^DateFilter"
If the user accepts the default values “10/1/2004” and “10/31/2004” as shown,
then the string “L 10/1/2004 U 10/31/2004” is assigned to the Replaceable
Parameter ^DateFilter and the second Run file command is executed as:
XRS FLUID.DAT VISC X S G $FILTER="DATE L 10/1/2004 U
10/31/2004"
This command creates X-bar and Standard Deviation charts of the variable VISC
in the file FLUID.DAT for all records where the value of the variable DATE is on
or between “10/1/2004” and “10/31/2004.”
Quality Analyst will make sure that the inputs meet the requirements for the
DateTime data type and that the first date is before (or the same as) the second.
AskAlphaSelectFilter
AskDateSelectFilter
AskIntegerSelectFilter
AskNumberSelectFilter
AskTimeSelectFilter
Note: See note at AskAlpha AskDate AskInteger AskNumber AskTime on
page 399.
The SelectFilter forms of the Ask command are intended to be used in conjunction
with a $FILTER parameter (see page 439) on a Run file command line that
requests a chart function.
This command displays a dialog box that prompts you for either one input (to use
as the criterion for a match filter) or two inputs (to use as the criteria for a range
filter), translates your entries into a valid argument for the $FILTER parameter,
and assigns the resulting string to a Replaceable Parameter. The syntax of the
SelectFilter form of the Ask command is:
command param "mainprompt" "prompt1" "prompt2" "prompt3"
"default1" "default2" "default3"
prompt1 the text to display in the dialog box to prompt the user for
the match input
prompt2 the text to display in the dialog box to prompt the user for
the first range input
prompt3 the text to display in the dialog box to prompt the user for
the second range input
default1 the value to display in the match input box when the dialog
box first appears
default2 the value to display in the first range input box when the
dialog box first appears
default3 the value to display in the second range input box when the
dialog box first appears
Use double quotation marks as shown in the syntax statement. All quoted
elements are optional from right to left. That is, you may omit a quoted element
only if you also omit all elements after it. You may, however, use empty quotation
marks for any of the quoted elements. In that case, standard text will be used for
prompts, and defaults will be left blank. To suppress any prompt, use a
quoted space. The default values, if specified, will be included in the string
assigned to the Replaceable Parameters when the user clicks Continue unless they
are first deleted or replaced.
If you omit the optional quoted parameters, this command displays the following
dialog box (varying slightly for the different data types):
Read more about Replaceable Parameters starting on page 484 and about
Filtering Data starting on page 90.
Example: The first of the following two Run file statements displays the dialog
box shown below.
AskDateSelectFilter ^DateFilter "Specify the Date(s) to
investigate:" "Day to investigate:" "Start of
investigation period:" "End of investigation
period:" "10/15/2004" "10/1/2004" "10/31/2004"
XRS FLUID.DAT VISC X S G $FILTER="DATE ^DateFilter"
If the user selects From and accepts the default values “10/1/2004” and
“10/31/2004” as shown, then the string “L 10/1/1999 U 10/31/2004” is assigned to
the Replaceable Parameter ^DateFilter and the second Run file command is
executed as:
XRS FLUID.DAT VISC X S G $FILTER="DATE L 10/1/2004 U
10/31/2004"
This command creates X-bar and Standard Deviation charts of the variable VISC
in the file FLUID.DAT for all records where the value of the variable DATE is on
or between “10/1/2004” and “10/31/2004.”
If the user selects Match and accepts the default value “10/15/2004” as shown,
then the string “M 10/15/2004” is assigned to the Replaceable Parameter
^DateFilter and the second Run file command is executed as:
XRS FLUID.DAT VISC X S G $FILTER="DATE M 10/15/2004"
This command creates X-bar and Standard Deviation charts of the variable VISC
in the file FLUID.DAT for all records where the value of the variable DATE is
“10/15/2004.”
Quality Analyst will make sure that the inputs meet the requirements for the
DateTime data type and that the first date in a range is before (or the same as) the
second.
AskDATFile
The AskDATFile command displays the Quality Analyst file-selection dialog box
and assigns the selected file name to a Replaceable Parameter. The syntax of the
AskDATFile command is:
AskDATFile param "path" "title"
AskList
The AskList command displays a dialog box that prompts you for a choice from a
drop down list of options and assigns that choice to a Replaceable Parameter. The
syntax of the AskList command is:
AskList param optlist "prompt" "default"
Use double quotation marks as shown in the syntax statement. Both prompt and
default are optional, but if you omit prompt, you must also omit default.
You can use empty quotation marks for prompt, in which case standard text will
be used. To suppress the prompt, use a quoted space. The default value, if
specified, will be assigned to the Replaceable Parameter when the user clicks
Continue unless it is first deleted or replaced.
An optlist is a set of possible values for the Replaceable Parameter that will
appear in the drop-down list. The list is enclosed in square brackets (“[ ]”) and the
options are enclosed in quotes and separated by fences (“|”). For example, this
optlist:
["ZN"|"PB"|"AS"]
If the possible values of the Replaceable Parameter are not sufficiently descriptive
to be useful to the user, you can include a “friendlier” string to display for each
option by adding a colon and a second quoted string to each option. For example,
this optlist:
["ZN":"Zinc"|"PB":"Lead"|"AS":"Arsenic"]
For any option in the list, the leftmost item will be assigned to the Replaceable
Parameter if that option is selected. In the example above, if the user selects
“Lead,” “PB” will be assigned to the Replaceable Parameter.
An optlist can contain both one-item and two-item options. For example, this
optlist:
["ZN"|"PB":"Lead"|"AS"]
Note that default, if included, must be the rightmost item of the desired default
option. For example, in this optlist:
["ZN"|"PB":"Lead"|"AS"]
Notice that because “AS” was specified as the default, “AS” is selected when
the dialog box appears.
Example 2: The following Run file command displays the dialog box shown
below.
AskList ^Element ["AS":"Arsenic"|"PB":"Lead"|"ZN":"Zinc"]
"Select an element to analyze:" "Lead"
Notice that “Lead” (the rightmost item of the option), not “PB” (the leftmost), is
correctly specified as the default.
AskOpenFile
The AskOpenFile command displays the Windows File Open dialog box and
assigns the selected file name to a Replaceable Parameter. The syntax of the
AskOpenFile command is:
AskOpenFile param "type" "path" "title"
Use double quotation marks as shown in the syntax statement. All quoted
elements are optional from right to left. That is, you may omit an element only if
you also omit all elements after it. You may, however, use empty quotation marks
for any of the quoted elements. Empty quotation marks for type will cause all file
types to be listed. Empty quotation marks for path will cause the contents of the
default Quality Analyst folder to be listed.
If you omit the optional elements, this command displays the following dialog
box:
AskSaveFile
The AskSaveFile command displays the Windows File Save dialog box and
assigns the selected file name to a Replaceable Parameter. The syntax of the
AskSaveFile command is:
AskSaveFile param "type" "path" "title"
Use double quotation marks as shown in the syntax statement. All quoted
elements are optional from right to left. That is, you may omit an element only if
you also omit all elements after it. You may, however, use empty quotation marks
for any of the quoted elements. Empty quotation marks for type will cause all file
types to be listed. Empty quotation marks for path will cause the contents of the
default Quality Analyst folder to be listed.
If you omit the optional elements, this command displays the following dialog
box:
AskVariable
The AskVariable command displays a dialog box that prompts you to select a
variable name from a list and assigns that variable name to a Replaceable
Parameter. The syntax of the AskVariable command is:
AskVariable param "filename" "type" "prompt" "COLUMN"
prompt the text to display in the dialog box to prompt the user for
input
COLUMN include "COLUMN" if you want the individual columns of
subgroup variables listed separately.
Use double quotation marks as shown in the syntax statement. Except for
filename, all quoted elements are optional from right to left. That is, you may
omit a quoted element only if you also omit all elements after it. You may,
however, use empty quotation marks for type or prompt. Empty quotation
marks for type will cause all variable types to be listed. Empty quotation marks
for prompt will cause a standard prompt to be displayed. To suppress the
prompt, use a quoted space.
The type string may include any character in the Type column of the table of
Data Types on page 37, except that instead of a number (for subgroup data), use a
space. "COLUMN" has no effect unless type contains a space.
If you omit the optional quoted parameters and specify WEIGHTS.DAT for
filename, this command displays the following dialog box:
Notice that only subgroup variables are listed because type is specified as " " (a
quoted space) in the statement.
Example 2: The following Run file statement displays the dialog box shown
below.
AskVariable ^Var1 "BOARDS.DAT" "P" "Please select a defect
column:"
Notice that only variables of data type “P” appear in the list.
ATTRIBUT
The ATTRIBUT command creates a Graphics file containing an attribute control
chart. The syntax of the ATTRIBUT command line in a Run file is:
ATTRIBUT inpfile var chart output [outfile]
Example: The command line to create a p-chart for the variable THREADS in
the Data file NOZZLE.DAT and produce the Graphics file THREADS.NWG is:
ATTRIBUT NOZZLE.DAT THREADS P G THREADS.NWG
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
BOXPLOT
The BOXPLOT command creates a Graphics file containing a box plot. The input
file for this command is an intermediate file created by an MPCAP command.
codes below.
type the type of box plot to create:
“B” for box & whiskers
“C” for capability box plot
“S” for ±3 sigma box plot
units the units for scaling plotted values:
“M” for measurement units
“S” for standardized units
specs “S” to display specifications and target
“N” to suppress them
output “G” to create a Graphics file
outfile the name of the output file, if applicable
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
CALC
The CALC command calculates calculated variables, if any, in a Data Set. The
syntax for the CALC command is:
CALC filename.dat
If your Data Set contains calculated variables and contains data that was added to
the Data Set outside the Quality Analyst Editor (for example, by other data-
collection software), you should include a CALC command in your Run file above
any charting or analysis commands.
Example: To create an Individual and Range chart for a variable in a Data Set
that contains a calculated variable, the Run file command lines would be:
CALC VENEER.DAT
IR VENEER.DAT THICK I R G
CHART
The CHART command displays an existing Graphics file on the screen. To use it,
place a CHART command line after one or more other command lines containing
Quality Analyst functions that create Graphics files. When used with the GROUP
command, CHART produces a multichart page layout for display. The CHART
command can also be used to create simultaneous on-screen and printed output,
and to save the chart with a new name or as a graphic image file.
CHART
Note that the default Graphics file name (GRAPHICS.NWG) need not be
specified if the CHART command follows immediately and only one chart is to be
displayed.
Example 2: Create the Graphics file WTPCAP.NWG, which contains a Process
Capability histogram of the variable WEIGHT in the file TEST.DAT, and display
it for 15 seconds:
PCAP TEST.DAT WEIGHT G WTPCAP.NWG
CHART WTPCAP.NWG /DELAY 15
Example 3: Create a 2×2 group layout of four p-charts from the variables
BUBBLES, MARKS, CRACKS, and TRIM in the file REJECTS.DAT. Note the
use of the four individual Graphics files later combined in the CHART command:
ATTRIBUT REJECTS.DAT BUBBLES P G REJ1.NWG
ATTRIBUT REJECTS.DAT MARKS P G REJ2.NWG
ATTRIBUT REJECTS.DAT CRACKS P G REJ3.NWG
ATTRIBUT REJECTS.DAT TRIM P G REJ4.NWG
CHART GROUP 2V2H REJ1.NWG REJ2.NWG REJ3.NWG REJ4.NWG
Example 4: To also print the group layout from Example 3, modify the last line:
CHART GROUP 2V2H REJ1.NWG ... REJ4.NWG /PRINT
To print the group layout in landscape orientation, modify the last line:
CHART GROUP 2V2H REJ1.NWG ... REJ4.NWG /PRINT LANDSCAPE
CHITEST
The CHITEST command calculates one of four types of Chi-square statistics, each
of which is appropriate for a particular type of data. Refer to the section titled Chi-
square beginning on page 344 for more details about Chi-square statistics.
The syntax of the CHITEST command line to calculate Bartlett’s Chi-square
statistic in a Run file is:
CHITEST inpfile B sampvars degfree output outfile
The syntax of the CHITEST command line to calculate the Difference Among
Proportions Chi-square statistic in a Run file is:
CHITEST inpfile D posresp sampsize output outfile
The syntax of the CHITEST command line to calculate the Equal Expected
Frequencies Chi-square statistic in a Run file is:
CHITEST inpfile E obsfreq output outfile
The syntax of the CHITEST command line to calculate the Unequal Expected
Frequencies Chi-square statistic in a Run file is:
CHITEST inpfile U obsfreq expfreq output outfile
Example 3: The command line to calculate the Equal Expected Frequencies Chi-
square statistic where, in the file BINCHK.DAT, the variable COUNT contains
observed frequencies, is:
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
CLEANUP
With the CLEANUP Run file command, you can convert files from external
sources (other software or systems) to Quality Analyst format or vice versa.
The parameters charlist and numcols are explained in more detail under
Importing a Comma-Separated-Value File on page 510.
Example: To convert the external file COMM6.DTX into the Quality Analyst
file CTEST.DAT, removing no characters, using the first row to determine the
number of columns, the Run file command line would be:
Before performing a fixed-field conversion from within a Run file, you must
create a Fixed-Field Definition file, as explained under Importing a Fixed-Field
File on page 512.
CLEAR
The CLEAR command clears, or “unassigns,” a Replaceable Parameter. The
syntax of the CLEAR command is:
CLEAR param
$COMMENT
You can add a line of text to the header lines of a chart by including a
$COMMENT command element and specifying the text as part of the chart
command. This is particularly useful in describing any filters that have been
applied, since Quality Analyst does not do this automatically in Run files. The
syntax is:
$COMMENT= "comment text"
The $COMMENT and $FILTER commands must come after all other parameters
of the main command, but they can occur in any order relative to each other.
Example: To add the comment “Lot Number 2250 Only” to a chart command
that includes a filter specification, include this line in your Run file:
IR FLUID.DAT VISC I R G $FILTER="LOT M 2250" $COMMENT="Lot
Number 2250 Only"
CONNECT
The CONNECT command is used by the Database Connectivity version of
Quality Analyst to establish a connection to an external database. See Using
Database Connectivity in Run files on page 583 for details.
The syntax for the CONNECT command is:
CONNECT filename.dat
COPY
The COPY command makes a copy of a file. The syntax of the COPY command
line in a Run file is:
COPY srcfile newfile
Both arguments should include the path if necessary to locate the source or the
destination.
Example: The command lines to make a temporary copy of the ELEMENTS
Data Set are:
COPY ELEMENTS.DAT ELEMTEMP.DAT
COPY ELEMENTS.NWH ELEMTEMP.NWH
CORVAR
The CORVAR command computes the linear correlation coefficients for two or
more variables and produces a correlation matrix for all possible combinations of
the selected variables. The syntax of the CORVAR command line in a Run file is:
CORVAR inpfile vars missing output outfile
Refer to the section titled Correlation on page 349 for a detailed explanation of
casewise and pairwise treatment of missing data.
Example: The command line to compute the correlation coefficients for the
variables PB, CD, ZN, HG, CR, MO, MN, and AS in the file ELEMENTS.DAT,
treating missing data casewise, and storing the results in a formatted text file is:
CORVAR ELEMENTS.DAT PB CD ZN HG CR MO MN AS C W
DEFAULT.NWG
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
CreateNWH
The CreateNWH command lets you create a Quality Analyst Header file from
within a Run file. (See Header Files on page 9 for details.) The syntax of the
CreateNWH command line in a Run file is:
CreateNWH hfile var1 type1 var2 type2 … varn typen
OR
Example 2: In this example, the effect is the same but the individual column
specifications are in a text file and the Run file command specifies that file using
the “@” form of the command.
CreateNWH PinWt.nwh @NewHeader.tmp
Notes: To specify other Data Set parameters stored in a Header file, use the
UpdateNWH command, described on page 478.
See Creating Header Files from within Run Files on page 495 for
additional information.
If you need to create a Header file that is compatible with a version of
Quality Analyst earlier than 6.1, make sure hfile uses the older “.hed”
file-name extension. Use “.nwh” to create a Header file in the newer
XML format.
CUSUM
The CUSUM command creates a Graphics file that contains a Cumulative Sum
chart. The syntax of the CUSUM command line in a Run file is:
CUSUM inpfile var vgroup output [outfile]
Example: The command line to chart the variable THICK from the Data file
VENEER.DAT and produce the Graphics file VENEER.NWG with the V-mask
located on the 22nd subgroup is:
CUSUM VENEER.DAT THICK 22 G VENEER.NWG
With Quality Analyst, you can produce a series of CUSUM charts with the
V-mask at different positions by creating a Run file containing multiple CUSUM
commands with the same variable.
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
DELETE
The DELETE command deletes a file. The syntax of the DELETE command line
in a Run file is:
DELETE filename
where filename is the name of the file to delete. Include the path to the file if
necessary to specify it.
Example: The command lines to delete a temporary copy of the ELEMENTS
Data Set might be:
DELETE ELEMTEMP.DAT
DELETE ELEMTEMP.NWH
EWMA
The EWMA command creates a Graphics file that contains an Exponentially
Weighted Moving Average chart (optionally in combination with a Range or
Standard Deviation chart). The syntax of the EWMA command line in a Run file
is:
EWMA inpfile var chart1 chart2 output [outfile]
Example 1: To create an EWMA chart for the variable THICK in the file
VENEER.DAT, and produce a Graphics file named VENEER.NWG for later use
by the PRINT or CHART commands, the Run file command line would be:
EWMA VENEER.DAT THICK E N G VENEER.NWG
Note the mandatory “N” (for “none”) for the nonexistent second chart type.
Example 2: To create an EWMA and Range chart for the same variable, and
produce a Graphics file for later use, the Run file command line would be:
EWMA VENEER.DAT THICK E R G
In this example, the absence of the outfile parameter tells Quality Analyst to
create a Graphics file with the default name GRAPHICS.NWG.
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
EXCEPT
The EXCEPT command generates an exception report which lists information
about data that produces control-limit, rule, or specification violations. (See
Exception Reporting on page 157 for details.)
EXCEPT qxp1 [qxp2 [...qxpn]] output [outfile]
EXTRACT
With the EXTRACT Run file command, you can extract subsets from Data files
and create a new file containing the selected data. The syntax for the Run file
command line is:
EXTRACT inpfile selfile outfile
Before executing the EXTRACT Run file command, you must create an Extract
Definition (SEL) file containing the information needed by the EXTRACT
command. Creating this file is explained under Extracting Subsets from a Data
File on page 536.
Example: To generate a file named LTNEW.DAT from the Data file
LTEX.DAT, using the Extract Definition file LTEX.SEL, the Run file command
line would be:
EXTRACT LTEX.DAT LTEX.SEL LTNEW.DAT
$FILTER
The Data Filter feature of Quality Analyst can be used in Run files to select
subsets of data. To do this, place the $FILTER command parameter at the end of a
line in the Run file that requests a chart function. You can include up to three
filters in the same chart command. The syntax is:
$FILTER="var type value [type value]" [$FILTER="..."]
[$FILTER="..."]
With the filter criterion, you can select data that either match a value (with “M”) or
fall within a range (with “L” and/or “U”). Only one boundary of a range need be
declared. Normally, the boundary values are included in the range. To exclude the
boundary values from the range, put a “>” in front of the lower limit and a “<” in
front of the upper limit.
In a single $FILTER command, the second type and value are allowable only
when you are specifying a second limit for values of the variable—that is, an
upper limit in addition to a lower limit.
The $FILTER and $COMMENT commands must come after all other parameters
of the main command, but they can occur in any order relative to each other.
Example 1: To request an Individual-and-Range chart for the variable VISC in
the file FLUID.DAT for Lot Number 2250 only (variable name LOT), include this
line in your Run file:
IR FLUID.DAT VISC I R G $FILTER="LOT M 2250"
Example 2: To request an X-bar and Standard Deviation chart for the variable
VISC in the file FLUID.DAT using only data from July 1995 (using the DATE
variable) where the value of the variable PH is greater than 7, include this line in
your Run file:
Without the “>” in the example above, the analysis would include data where the
value of PH is greater than or equal to 7.0.
Note: To prompt users for filter criteria, use any of the Run file commands that
begin with “Ask” and end with “Filter” such as AskDateRangeFilter.
FLIP
With the FLIP Run file command, you can rotate Data files 90 degrees, turning
rows into columns and columns into rows. To rotate rows, the syntax of the Run
file command line is:
FLIP inpfile R rows D outfile
To rotate columns, the syntax of the Run file command line is:
FLIP inpfile C cols D outfile
GROUP
The GROUP command is really an optional parameter to the CHART and PRINT
commands that instructs the Run File Interpreter to form a multichart Group
Layout in a specified arrangement or “style.” The syntax of the GROUP parameter
in a Run file command line is:
The following table lists the available layout styles and their command-line codes.
2V 3V 4V 5V
The section Group Layout on page 196 discusses creating Group Layouts
interactively with the Graphics Viewer.
Example: To print two smaller charts (SAMPLE1.NWG and SAMPLE2.NWG)
over a third full-width chart (SAMPLE3.NWG), use style code “2VST”:
PRINT GROUP 2VST SAMPLE1.NWG SAMPLE2.NWG SAMPLE3.NWG
Note that the GROUP command option and the style code follow the CHART or
PRINT command immediately and come before the file names. The charts from
the specified files are placed in the specified layout left-to-right then top-to-
bottom.
IR
The IR command creates a Graphics file containing an Individuals (and/or a
Range) chart. The syntax of the IR command line in a Run file is:
IR inpfile var chart1 chart2 output [outfile]
Example 1: To create an Individual chart for the variable THICK in the file
VENEER.DAT, and produce a Graphics file named VENEER.NWG for later use
by the PRINT or CHART commands, the Run file command line would be:
IR VENEER.DAT THICK I N G VENEER.NWG
Note the mandatory “N” (for “none”) for the nonexistent second chart type.
Example 2: To create an Individual and Range chart for the same variable, and
produce a Graphics file for later use, the Run file command line would be:
IR VENEER.DAT THICK I R G
In this example, the absence of the outfile parameter tells Quality Analyst to
create a Graphics file with the default name GRAPHICS.NWG.
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
KRUSKAL
The KRUSKAL command performs a Kruskal-Wallis test, which is a
nonparametric analysis of variance for populations where the assumptions
required by one-way ANOVA are not met. The syntax of the KRUSKAL
command line in a Run file is:
KRUSKAL inpfile vars output outfile
Example: The command line to perform the Kruskal-Wallis test on the variables
SINKS, MOUNT, MARKS, TRIM, and EDGE in the file DEFECTS.DAT is:
KRUSKAL DEFECTS.DAT SINKS MOUNT MARKS TRIM EDGE W
DEFAULT.NWG
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
MCAPREP
The MCAPREP command creates a formatted text file containing a process
capability report. The input file for this command is an intermediate file created by
an MPCAP command.
The syntax of the MCAPREP command in a Quality Analyst Run file is:
MCAPREP inpfile statlist output outfile
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
Example 2: To create the same report in the form of a Data file named
CAPREP1.DAT, the command line is:
MCAPREP MPCAP.IDF 9 31 32 38 49 D CAPREP1.DAT
MERGE
With the MERGE Run file command, you can insert, delete, or extract rows or
columns from a file. You can also extract or replace an entire block from within a
file and merge files in any orientation.
You specify the MERGE operation by including an operation number in the
command line with other parameters appropriate for the operation. The following
table lists the operation numbers, the operation, and the command syntax for that
operation.
Operation
Operation Command Syntax
Number
1 Extract Rows MERGE inpfile 1 begrow endrow outfile
2 Extract MERGE inpfile 2 begcol endcol outfile
Columns
3 Extract a MERGE inpfile 3 begrow endrow begcol endcol outfile
Block
4 Delete Rows MERGE inpfile 4 begrow endrow output [outfile]
5 Delete MERGE inpfile 5 begcol endcol output [outfile]
Columns
6 Insert Rows MERGE inpfile 6 insfile begrow output [outfile]
7 Insert MERGE inpfile 7 insfile begcol output [outfile]
Columns
8 Replace a MERGE inpfile 8 replfile begrow begcol output
Block [outfile]
MPCAP
The MPCAP command prepares an intermediate file for use by the BOXPLOT
(page 423) and MCAPREP (page 443) Run file commands. The syntax of the
MPCAP command line in a Run file is:
MPCAP inpfile vars bkdntype [bkdncol] [bkdnsort] [bkdnint]
output outfile
The “A” option for output lets you execute several MPCAP commands on
different Data files and include them all in a single report. Be sure to use the “C”
option on the first MPCAP command in the series.
Example 1: The following command line creates an intermediate file named
DXSHIFT.TMP containing process capability information on the variable
DIAMETER in the file DIAM.DAT. The information is broken down and sorted
by the variable SHIFT.
MPCAP DIAM.DAT DIAMETER B SHIFT S C DXSHIFT.TMP
MR
The MR command creates a Graphics file containing a Median or
Median/Individual chart (and/or a Range chart). The syntax of the MR command
line in a Run file is:
MR inpfile var chart1 chart2 output [outfile]
Example 1: To create a Median chart for the variable THICK in the file
VENEER.DAT, and produce a Graphics file named VENEER.NWG for later use
by the PRINT or CHART commands, the Run file command line would be:
MR VENEER.DAT THICK M N G VENEER.NWG
Note the mandatory “N” (for “none”) for the nonexistent second chart type.
Example 2: To create a Median/Individuals and a Range chart for the same
variable, and produce a Graphics file for later use, the Run file command line
would be:
MR VENEER.DAT THICK MI R G
In this example, the absence of the outfile parameter tells Quality Analyst to
create a Graphics file with the default name GRAPHICS.NWG.
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
MSORT
With the MSORT (for multikey sort) Run file command, you can sort a Data file
based on one or more columns of alpha, numeric, time, or DateTime data. You can
sort in either ascending or descending order and you can either replace the original
data with the sorted data or send the sorted data to a new file. The syntax of the
MSORT command line in a Run file is:
MSORT inpfile sortvars sortorder output [outfile]
Example: To generate a new file named LTSORT.DAT from the Data file
LTEX.DAT, sorted in ascending order by the variable “DATE,” the Run file
command line would be:
MSORT LTEX.DAT DATE A C LTSORT.DAT
MULTREG
With the MULTREG Run file command, you can perform a multiple linear
regression. The syntax of the MULTREG command line in a Run file is:
MULTREG inpfile depvar indvars confint resid output
[outfile]
For confint, specify “C” to include the confidence intervals at 95% based on the
individual measurements, “N” for no confidence intervals, or a number from 1 to
99 to specify the probability level percent. If you specify a number, follow it with
either “I” to base confidence intervals on the individual measurements, or “E” to
base confidence intervals on the equation. You must specify confidence intervals
even if you do not select graphics output.
Example: To perform a multiple linear regression on the dependent variable
TENSILE and the independent variables MIX, BINDER, MOISTURE, and
DWELL in the file REGTEST.DAT, and send output that includes residuals to a
Graphics file, the Run file command line would be:
MULTREG REGTEST.DAT TENSILE MIX BINDER MOISTURE DWELL C R
G
The “G” and “W” options create Graphics files but do not display or print them.
This is accomplished by an additional command line containing a “CHART” or a
“PRINT” command or both.
NEWVAR
With the NEWVAR Run file command, you can create new variables in a Data
file from calculations on existing data in the file. The syntax of the NEWVAR
command line in a Run file is:
NEWVAR inpfile vdffile output [outfile]
Before executing the NEWVAR Run file command, you must create a Variable
Definition (VDF) file containing the information needed by the NEWVAR
command. Creating this file is explained under Defining Calculations for Run
files on page 538.
Example: To generate a new file named LTNEW.DAT from the Data file
LTEX.DAT, using the Variable Definition file LTEX.VDF, the Run file command
line would be:
NEWVAR LTEX.DAT LTEX.VDF C LTNEW.DAT
NORMPLOT
With the NORMPLOT Run file command, you can create a normal probability
plot for a single variable or a quantile plot for two variables.
To create a normal probability plot, the syntax of the Run file command line is:
NORMPLOT inpfile N var output [outfile]
To create a quantile plot, the syntax of the Run file command line is:
NORMPLOT inpfile Q xvar yvar output [outfile]
Example: To produce a normal probability plot using the variable THICK from
the file VENEER.DAT, and create a graphics output file with the default file
name, the Run file command line is:
NORMPLOT VENEER.DAT N THICK G
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
NoWorkingMessage
The NoWorkingMessage command tells the Run File Interpreter to stop
displaying the “Executing” message box while a Run file is running.
This command can only be used to turn the message off, not on. Its effect lasts
until the Run file terminates. It has no effect on the message boxes displayed by
the PAUSE command.
PARETO
The PARETO command creates a Graphics file containing a Pareto diagram. The
PARETO command can accept either a multi-variable file or a single-variable file
and allows charts with multiple variables (multiple defects). The syntax of the
PARETO command line in a Run file is:
PARETO inpfile format vars output [outfile]
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
ParseFileName
The ParseFileName command lets you extract the path, file name, and extension
from a full file path. The parsed components are assigned to Replaceable
Parameters. The syntax of the ParseFileName command line in a Run file is:
ParseFileName path base ext "fullname"
PAUSE
The PAUSE command pauses the Run file. The syntax is:
PAUSE [seconds] [message]
On this (and all other PAUSE dialog boxes), select Continue button to resume the
Run file, Quit to terminate the execution of the innermost Run file, if it is nested,
or Quit All to terminate the execution of all Run files including any that are nested.
If the PAUSE command includes a seconds parameter, the Run file will pause
for the specified number of seconds and then continue automatically. The PAUSE
dialog box will include a “timer.”
One of the hands on the timer will move each second. Select Continue to resume
the Run file before the programmed pause expires.
For example, to pause for 20 seconds and then continue automatically, the Run
file command line would be:
PAUSE 20
If the PAUSE command line includes a message parameter, the PAUSE dialog
box will display the message you specify. For example, this Run file command:
PAUSE Get your printer ready, then select Continue.
PCAP
The PCAP command creates a Graphics file containing a process capability
histogram. The syntax of the PCAP command line in a Run file is:
PCAP inpfile var output [outfile]
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
POLREG
With the POLREG Run file command, you can fit your data, using the least-
squares method, to a polynomial equation of the form:
y = B0 + B1x + B2x2 + B3x3 + ... Bnxn
where y is the dependent variable and x is the independent variable. To perform a
polynomial regression, the syntax for the Run file command line is:
POLREG inpfile indvar depvar order confint resid output
[outfile]
For confint, specify “C” to include the confidence intervals at 95% based on the
individual measurements, “N” for no confidence intervals, or a number from 1 to
99 to specify the probability level percent. If you specify a number, follow it with
either “I” to base confidence intervals on the individual measurements, or “E” to
base confidence intervals on the equation. You must specify confidence intervals
even if you do not select graphics output.
If you specify graphics output and residuals output, the resulting Graphics file will
contain the residuals plot, not the regression plot. If you want the Graphics file to
contain the regression plot, do not specify residuals output. If you want to generate
Graphics files for both the residuals and the regression, include separate POLREG
commands for each in your Run file.
Example: To perform a third-order polynomial regression with the independent
variable LOAD and dependent variable RESPONSE in the file POLYTEST.DAT,
with the regression plot in the graphics output file, the Run file command line
would be:
POLREG POLYTEST.DAT LOAD RESPONSE 3 C N G
The “G” and “W” options create Graphics files but do not display or print them.
This is accomplished by an additional command line containing a “CHART” or a
“PRINT” command or both.
PRINT
The PRINT command sends output from Graphics files to the default Windows
printer without displaying the graph on the screen. To use it, place a PRINT
command line after one or more other command lines containing Quality Analyst
functions that create Graphics files. When used with the GROUP command,
PRINT prints a multichart page layout.
PRINT Command
Description
Options
GROUP style Instructs PRINT to form a multichart layout in the specified
style. Style is followed by a list of Graphics files to
include in the Group Layout. These Graphics files must all
have been produced before this command is executed. See
GROUP on page 440, for examples and a list of layout
styles.
NONE Used in place of a Graphics file name following a GROUP
command to leave part of a Group Layout blank.
/LANDSCAPE Prints the graphics in landscape orientation. (The default is
portrait.)
Note that the default Graphics file name (GRAPHICS.NWG) need not be
specified if the PRINT command follows immediately and only one chart is to be
displayed.
Example 2: Create the Graphics file CAPX.NWG, which contains a Process
Capability histogram of the variable WEIGHT in the file TEST.DAT, and print it
in landscape orientation:
PCAP TEST.DAT WEIGHT G CAPX.NWG
PRINT CAPX.NWG /LANDSCAPE
Example 3: Print a 2×2 group layout of four p-charts from the variables
BUBBLES, MARKS, CRACKS, and TRIM in the file REJECTS.DAT.
ATTRIBUT REJECTS.DAT BUBBLES P G REJ1.NWG
PROMPT
Note: The PROMPT command has been superseded by the newer and more
flexible “Ask” commands which are described beginning on page 399.
With the PROMPT Run file command, you can have your Run file pause and
request input (such as a file name, variable name, text string, or filter criterion) to
assign to a Replaceable Parameter (see Using the “$” Replaceable Parameters on
page 488). You can include a message to give the user of the Run file specific
information about the item to be specified. To prompt for user input, the syntax for
the Run file command line is:
PROMPT parameter [message]
The PROMPT command will display a dialog box whose appearance depends on
the type of parameter requested. Here are a few examples.
This command line:
PROMPT $FILE1 Enter the name of the Data file to analyze:
During the execution of the Run file, you simply enter, when prompted, the
requested value(s) and click Continue to proceed to the next Run file command. If
you select Quit, the execution of the Run file (whose name is displayed in the
recessed box labeled Run file:) will stop. If the PROMPT command is in a nested
Run file, only the innermost Run file will stop.
For an explanation of what the Run File Interpreter does with the values you enter,
see Replaceable Parameters beginning on page 484.
REDUCE
With the REDUCE Run file command, you convert a Data file containing
measurement data in subgroups into a Data file contain the average (mean or
X-bar), range, and standard deviation of each subgroup. The syntax for the Run
file command line is:
REDUCE inpfile var output [outfile]
Quality Analyst will automatically include description variables from the input file
in the output file.
Example: To reduce the subgrouped variable WEIGHT in the file
TOMATO.DAT, and send the output to a new Data file named
TOMATXRS.DAT, the Run file command line would be:
REDUCE TOMATO.DAT WEIGHT C TOMATXRS.DAT
REGRESS
With the REGRESS Run file command, you can compute the least-squares
coefficients of a single variable regression for the following forms of equations:
Linear: y = A + Bx
Exponential: y = A + eBx
Logarithmic: y = A + B×Ln(x)
Power Law: y = A × xB
To perform a regression analysis, the syntax for the Run file command line is:
REGRESS inpfile indvar depvar types detail resid [confint]
output [outfile]
If you omit confint, the graph will include the confidence intervals at 95%
based on the individual measurements. Confidence intervals are not displayed if
more than one regression type is selected.
If you specify graphics output and residuals output, the resulting Graphics file will
contain the residuals plot, not the regression plot. If you want the Graphics file to
contain the regression plot, do not specify residuals output. If you want to generate
Graphics files for both the residuals and the regression, include separate
REGRESS commands for each in your Run file.
Example 1: To perform linear regression on the variables LEVEL and OUTPUT
in the file LOGDATA.DAT, with the regression plot in the graphics output file,
the Run file command line would be:
REGRESS LOGDATA.DAT LEVEL OUTPUT LIN N N G
Example 2: To perform the same analysis as in Example 1, but with the residuals
plot instead of the regression plot in the graphics output file, the Run file
command line would be:
REGRESS LOGDATA.DAT LEVEL OUTPUT LIN N R G
Example 3: To perform the same analysis as in Example 1, but with text output
(including details) in the file REG1.NWG, the Run file command line would be:
REGRESS LOGDATA.DAT LEVEL OUTPUT LIN D N W REG1.NWG
The “G” and “W” options create Graphics files but do not display or print them.
This is accomplished by an additional command line containing a “CHART” or a
“PRINT” command or both.
REM
With the REM Run file command, you can include comments in a Run file. This
lets you annotate the Run file and describe or explain other command lines. Any
line beginning with “REM” will be ignored during the execution of the Run file.
For example:
REM Create X-bar chart and print it:
XRS VENEER.DAT THICK X N G
PRINT
RENAME
The RENAME command renames a file. The syntax of the RENAME command
line in a Run file is:
RENAME oldname newname
Both arguments should include the path if necessary to specify the files.
Example: The command line to rename the default Graphics file
GRAPHICS.NWG so that it won’t be overwritten might be:
RENAME GRAPHICS.NWG WTXR.NWG
REPORT
With the REPORT Run file command, you can print all or selected columns of
Quality Analyst Data files in a format suitable for reports, summaries, and hard-
copy records. You define the format for the report in a format definition file.
The syntax for a REPORT Run file command line is:
REPORT inpfile rptfile output outfile [$PAGEWIDTH=n]
rptfile the name of the Report Format (RPT file (see below))
output “W” to create a Graphics file containing a text report
outfile the name of the output file
$PAGEWIDTH=n the number of characters per line of output
Before executing the REPORT Run file command, you must create a Report
Format file that contains the information needed by the REPORT command.
Creating this file is explained under Printing a Formatted Report on page 115.
You can control the width of a printed report by including a $PAGEWIDTH=n
parameter on the REPORT command line. Substitute the number of characters per
line of output (up to 240) for n (the default is 120).
Example: To generate a report from the Data file LTEX.DAT, using the format
definition file LTEX.RPT, and send the output to a formatted text file named
LTEX.NWG, the Run file command line would be:
REPORT LTEX.DAT LTEX.RPT W LTEX.NWG
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
RUN
With the RUN Run file command, you can “nest” Run files. The syntax of the
command line is:
RUN filename
where filename is the name of a Run file. For example, the following Run file
command line will execute the Run file named SECOND.RUN and resume with
the next command line in the original Run file.
RUN SECOND.RUN
You can use the PROMPT command to request the name of the nested Run file
from the user. For example:
PROMPT $FILE3 Enter the name of the secondary run file:
RUN "$FILE3"
In this example, enclosing “$FILE3” in quotes in the RUN command will avoid an
error if the user enters a long file name or one that contains a space.
One useful application of “nested” Run files is error recovery. If an error occurs in
a Run file, the Run file operation terminates. If an error occurs in a nested Run
file, the execution of the “parent” Run file resumes at the next command. (This is
the default behavior, which the user can set in Quality Analyst Settings.) This is
very handy if you anticipate errors (such as missing Data files or insufficient data
after filtering) and want the Run file to continue when those errors occur. In this
case, the portion of the Run file with the error potential should be “nested.”
RUNCHART
The RUNCHART command creates a Graphics file containing a run chart. The
syntax of the RUNCHART command line in a Run file is:
RUNCHART inpfile xvar yvars output [outfile]
Both the “G” and “M” options create a Graphics file. If you specify only one
Y variable, output should be “G”. If you specify more than one Y variable,
output type “G” produces a chart on which all Y variables are scaled to a single
Y axis (see example on page 168), and output type “M” produces charts on which
each Y variable is scaled to its own Y axis (see example on page 169).
Example 1: To plot the variable MOISTURE against the row number from the
Data file SAMPLE.DAT, and send output to the Graphics file SAMPLE.NWG,
the Run file command line would be:
RUNCHART SAMPLE.DAT A MOISTURE G SAMPLE.NWG
To plot them on separate Y axes, the Run file command line would be:
RUNCHART SAMPLE.DAT A MOISTURE PROTEIN WEIGHT M
The “G” and “M” options both create a Graphics file but do not display or print
them. This is accomplished by an additional command line containing either a
“CHART” or a “PRINT” command.
SAVE
The SAVE Run file command lets you save a Graphics file with a new name or as
a graphic image file in any of several common file formats. It also lets you save
several Graphics files in a Group layout. The syntax for the SAVE command in a
Run file is:
SAVE inpfile AS outfile width height
OR
If outfile has the extension NWG, width and height are ignored and may be
omitted.
Example 1: To save the chart contained in LeadMR.NWG to an 800×600-pixel
graphic image file named LeadMR.jpg, the Run file command line would be:
SAVE LeadMR.NWG AS LeadMR.jpg 800 600
See also: the /SAVEAS parameter of the CHART command on page 426.
SCATTER
With the SCATTER Run file command, you can create an automatically-scaled
scatter diagram of one X-axis variable and up to eight Y-axis variables in a Data
file. The syntax for the SCATTER command line in a Run file is:
SCATTER inpfile xvar yvars output [outfile]
Example: To plot the variables PH and MOISTURE against the variable CONC
from the Data file SAMPLE.DAT, and send output to the Graphics file
SAMPLE.NWG, the Run file command line would be:
SCATTER SAMPLE.DAT CONC PH MOISTURE G SAMPLE.NWG
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
SET
The SET command assigns a value to a Replaceable Parameter. The syntax for the
SET command is:
SET param = "value"
the first two assign “WEIGHT.DAT” and “WT” to the Replaceable Parameters
“^File1” and “^Var1” respectively so that the third executes as:
XRS WEIGHT.DAT WT X R G
The following examples illustrate situations where using the SET command
(rather than “hard coding” the parameters) is particularly useful.
Example 2: Creating and maintaining a Run file with repeated commands can be
easier using the SET command. In this example the ATTRIBUT and CHART
commands are all identical but they produce different results because of the SET
commands.
SET ^InFile = "DEFECTS.DAT"
SET ^Var = "FOREIGN"
ATTRIBUT ^InFile ^Var U G
CHART
SET ^Var = "BUBBLES"
ATTRIBUT ^InFile ^Var U G
CHART
SET ^Var = "SINKS"
ATTRIBUT ^InFile ^Var U G
CHART
SET ^Var = "MOUNT"
ATTRIBUT ^InFile ^Var U G
CHART
SET ^Var = "MARKS"
ATTRIBUT ^InFile ^Var U G
CHART
SetCurrentDir
The SetCurrentDir Run file command lets you specify a file path that the Run File
Interpreter will use as the current directory with file names that do not have paths
specified. The syntax is:
SetCurrentDir "path"
where path is the file path. The path should be enclosed in double quotes,
especially if the path includes folder names that are long or contain spaces.
Examples:
SetCurrentDir "G:\LSmith\QA\Data"
SetCurrentDir "D:\Quality Analyst\Data"
SetCurrentDir "S:\QA\DATA"
SHELL
With the SHELL Run file command, you can execute a DOS or Windows
command from within a Run file as if from a Windows command line or a DOS
system prompt. After the command executes, the Run file resumes with the next
command line. The syntax is:
SHELL command [parameters]
The SHELL command can be particularly useful if the purpose of your Run file is
to integrate Quality Analyst with other applications. The command you specify on
the SHELL command line can, among other things, call a batch file, or execute
another application.
Example: To run a third-party program named DATAMSTR.EXE, the command
line would be:
SHELL DATAMSTR.EXE
Use SHELL only if you are familiar with its operation. It will only work on
systems using DOS version 3.0 or higher.
SINGCOL
With the SINGCOL Run file command, you can convert a single column of data
into a Data file containing multiple columns of data, and vice versa.
To convert several data columns to a single column, the syntax for the command
line is:
SINGCOL inpfile T vars desc [descvars] D outfile
To convert from a single column of data into multiple columns, the syntax for the
command line is:
SINGCOL inpfile F var size desc [descvars] D outfile
Example: To convert the column WEIGHT in the input file TEST1.DAT to four
columns in the output file TEST2.DAT, and include the description variables
DATE and LOTNUM, the Run file command line would be:
SINGCOL TEST1.DAT F WEIGHT 4 D DATE LOTNUM D TEST2.DAT
SUMMARY
Note: The SUMMARY command has been superseded by the newer EXCEPT
command (page 437) which is more flexible and capable.
With the SUMMARY Run file command, you can summarize the output of a
Quality Analyst chart command. You can generate either a list of out-of-control
points in the file, or a summary or tabulation of out-of-control points based on
user-selected description fields. To generate a summary report, the syntax of the
Run file command line is:
SUMMARY inpfile type output outfile
The inpfile for a SUMMARY command is the output file from a control chart
command with the “D” output option. You must therefore precede a SUMMARY
command with a chart command that includes the “D” output option and a
Summary (SMY) file name.
Example: To generate both list output and summary output from the summary
file VENEER.SMY, and send the output to a formatted text file named
VENEER.NWG, the Run file command line would be:
SUMMARY VENEER.SMY B W VENEER.NWG
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
SUMSTAT
With the SUMSTAT Run file command, you can compute “sample” or
“population” descriptive statistics for one or more variables in a Data file. To
compute descriptive statistics, the syntax for the Run file command line is:
SUMSTAT inpfile vars type output outfile
Example: To calculate “sample” statistics for the variable WEIGHT in the file
TOMATO.DAT, and send output to a formatted text file named TOMATO.NWG,
the Run file command line would be:
SUMSTAT TOMATO.DAT WEIGHT S W TOMATO.NWG
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
SYSTEM
The SYSTEM command stops the execution of the Run file and the Run File
Interpreter, returning you to Windows. The syntax is:
SYSTEM
TITLE
With the TITLE Run file command, you define a title for the Run file. Any text
following this command will be used as the title. This command must be the first
line in the Run file. For example:
TITLE Northwest Manufacturing Co.
Run file titles are displayed next to the Run file names in the Execute Run file
dialog box.
TTESTS
The TTESTS command calculates one of four types of t-statistics. Refer to the
section titled t-Statistics beginning on page 356 for a detailed explanation of
t-statistics.
The syntax of the TTESTS command line to calculate an ordinary unpaired
sample t-statistic in a Run file is:
TTESTS inpfile U var1 var2 samevari hypmeandiff output
outfile
The syntax of the TTESTS command line to calculate a paired sample t-statistic
in a Run file is:
TTESTS inpfile P var1 var2 hypmeandiff output outfile
The syntax of the TTESTS command line to calculate a single sample t-statistic
in a Run file is:
TTESTS inpfile S var hyppopmean popvari output outfile
The syntax of the TTESTS command line to calculate the Pearson Product
Moment Correlation in a Run file is:
TTESTS inpfile M var1 var2 output outfile
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
UpdateNWH
The UpdateNWH command lets you modify certain parameters in a Quality
Analyst Header file from within a Run file. (See Header Files on page 9 for
details.) The syntax of the UpdateNWH command line in a Run file is:
UpdateNWH hfile var param value
OR
Example 2: In this example, the effect is the same but the individual command
elements are in a text file and the Run file command specifies that file using the
“@” form of the command.
UPDATENWH PIN.NWH @UpdParam.tmp
Note: To create a Header file from within a Run file use the CreateNWH
command, described on page 433.
See Creating Header Files from within Run Files on page 495 for
additional information.
If you need to update a Header file that is compatible with a version of
Quality Analyst earlier than 6.1, make sure hfile uses the older “.hed”
file-name extension. Use “.nwh” to update a Header file in the newer
XML format.
UpdateNWH replaces UpdateHED which was used by Quality Analyst
before version 6.1. For backwards compatibility, both commands are
identical.
USETEMPDIR
The USETEMPDIR command creates a temporary directory (folder) on your local
PC and makes that directory the current Quality Analyst directory. The syntax for
the USETEMPDIR command line in a Run file is:
USETEMPDIR [ON]|OFF
Note If the file name in the CONNECT command or any charting command
includes the full path to the Data file, Quality Analyst will use the full
path, overriding USETEMPDIR.
Example: The Run file listed below and the Data Set EXTDATA.DAT and
EXTDATA.NWH are located in a shared directory, G:\QA\DATA\.
USETEMPDIR ON
CONNECT EXTDATA.DAT
PCAP EXTDATA.DAT VAR G 1.NWG
CHART 1.NWG
4. Open
C:\WINDOWS\TEMP\QA\EXTDATA.NWH and
retrieve external data into
C:\WINDOWS\TEMP\QA\EXTDATA.DAT.
PCAP EXTDATA.DAT VAR 5. Create a Graphics file
G 1.NWG (C:\WINDOWS\TEMP\1.NWG‡) containing a
process-capability histogram of the variable VAR
from C:\WINDOWS\TEMP\QA\EXTDATA.DAT.
CHART 1.NWG 6. Display the chart in the file
C:\WINDOWS\TEMP\1.NWG.
[Run file ends] 7. Delete all files in C:\WINDOWS\TEMP\QA\ then
delete the directory.
† Assuming the Quality Analyst Temporary File Path is C:\WINDOWS\TEMP\.
‡ Assuming the Quality Analyst Graphics and Text File Path is C:\WINDOWS\TEMP\.
See File Paths and Permissions on page 269 for details.
WALDRUN
The WALDRUN command performs a Wald-Wolfowitz test to determine the
degree of randomness in a data series. Refer to the section titled Wald-Wolfowitz
Runs Test on page 364 for a detailed explanation of the Wald-Wolfowitz test.
The syntax of the WALDRUN command line in a Run file is:
WALDRUN inpfile var conditions output outfile
For conditions, enter two conditions for dividing your data into two mutually
exclusive sets. Each condition consists of a comparison operator (from the table
below) and a value. For example, to divide your data into sets containing values
that are less than 134 and greater than 134, condition would be:
Comparison Operators
Operator Meaning
.LT. less than
.GT. greater than
.LE. less than or equal to
.GE. greater than or equal to
Example: The command line to perform the Wald-Wolfowitz test on the variable
VISCOS in the file NP7.DAT where values of VISCOS are divided into those less
than 1300 and those greater than or equal to 1300, is:
WALDRUN NP7.DAT VISCOS .LT. 1300 .GE. 1300 W DEFAULT.NWG
The “W” option creates a Graphics file but does not display or print it. This is
accomplished by an additional command line containing either a “CHART” or a
“PRINT” command.
WEIBULL
With the WEIBULL Run file command, you can fit the Weibull or the exponential
probability distribution to data and produce a probability plot. The syntax for the
Run file command line is:
WEIBULL inpfile var output [outfile]
Example: To use the variable LIFE from the file FAN.DAT, and create a
graphics output file, the Run file command would be:
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
XRS
The XRS command creates a Graphics file containing an X-bar, X-bar and Range,
or X-bar and Standard Deviation chart. The syntax of the XRS command line in a
Run file is:
XRS inpfile var chart1 chart2 output [outfile]
Example 1: To create an X-bar chart for the variable THICK in the file
VENEER.DAT, and produce a Graphics file named VENEER.NWG for later use
by the PRINT or CHART commands, the Run file command line would be:
XRS VENEER.DAT THICK X N G VENEER.NWG
Note the mandatory “N” (for “none”) for the nonexistent second chart type.
Example 2: To create an X-bar and Range chart for the same variable, and
produce a Graphics file for later use, the Run file command line would be:
XRS VENEER.DAT THICK X R G
In this example, the absence of the (outfile) parameter tells Quality Analyst to
create a Graphics file with the default name GRAPHICS.NWG.
The “G” option creates a Graphics file but does not display or print the graphic.
This is accomplished by an additional command line containing either a “CHART”
or a “PRINT” command.
Replaceable Parameters
Introduction
A Quality Analyst Run file may contain Replaceable Parameters wherever certain
parameters (such as file names, variable names, or filter criteria) are normally
expected.
Once a user has assigned a value to a Replaceable Parameter, the Run File
Interpreter will substitute that value for each occurrence of the parameter in the
Run file.
The following discussion mainly addresses the “^” Replaceable Parameters. The
“$” Replaceable Parameters are discussed under Using the “$” Replaceable
Parameters beginning on page 488.
To assign a value to a Replaceable Parameter on the command line for the Run
File Interpreter you must first include a Replaceable Parameter in the Run file. For
example, let’s say the Run file WTTASK.RUN contains the following statement.
XRS ^File1 WT X R G
If the Run file that contains this line is executed from this command line:
QARFI.EXE $RUN=WTTASK.RUN ^File1="Ring Weight.dat"
The Run File Interpreter will try to interpret “Ring” as the file name and
“Weight.dat” as the variable and the remaining parameters in the statement will
be off by one position. This will produce an error message.
To avoid this, place the Replaceable Parameter in quotes:
XRS "^File1" WT X R G
In this case, that same Run file command line will cause this statement to be
interpreted as:
XRS "Ring Weight.dat" WT X R G
and it will execute correctly because Quality Analyst will interpret everything
within the quotes as the file name.
The FILTER parameters let you specify the [type] and [value] elements of
the $FILTER command option (described on page 439) at run time. (See the
example below.)
Quality Analyst lets a single Run file contains up to 35 “$” Replaceable
Parameters of each type by allowing the “n” in each parameter name to be a digit
from “1” to “9” or a letter from “A” to “Z.” Legal Replaceable Parameters
include $FILE1, $VAR9, $STRINGA, and $RANGEFILTERZ.
Note: The $STRINGn parameter accepts all text entered by the Run file user as
is, and is therefore potentially troublesome. The Run file author is well-
advised to include explicit instructions in the message portion of the
PROMPT command when including a $STRINGn Replaceable
Parameter.
You can assign values to “$” Replaceable Parameters in one of two ways:
Include a PROMPT command for each Replaceable Parameter in the Run file.
When the PROMPT command is executed, the user enters the “value” of the
Replaceable Parameter (like a file name) in the Prompt dialog box. (See
PROMPT on page 460). You will probably find this method most useful.
Specify the values of the Replaceable Parameters in the Windows command
line when you execute the Run file.
Example: During the execution of a Run file, if the user is asked for a Data file
name with the command:
PROMPT $FILE3 Enter a file name to analyze:
and the user enters COOKIE.DAT when prompted, then the Run file command:
XRS $FILE3 WEIGHT X R G
In the same manner, users can be prompted for variable names, filter criteria, or
text strings (for use as labels or special command parameters). For example, if the
user makes the following assignments at run time (either on the command line or
in response to PROMPT commands):
Note: If the Run File Interpreter encounters a “$” Replaceable Parameter for
which a value has not been provided, the user will be prompted for the
value of the parameter as if a PROMPT command without a message
parameter had been executed.
Example 1
In this example, the Run file will create an X-bar and Range chart on the screen
for the variable THICK in the Data file VENEER.DAT. It will then create a
process-capability distribution on the screen. The Run file will pause when each
chart is displayed and wait for the user to resume execution. The default Graphics
file, GRAPHICS.NWG, will be used since no file name is specified. The Run file
contains the following lines:
TITLE XR and PCAP of veneer thickness
XRS VENEER.DAT THICK X R G
CHART
PCAP VENEER.DAT THICK G
CHART
Example 2
In this example, we expand the first example to add an Individuals chart and a
Normal Probability plot and then assemble the charts into a group layout for
printing.
Example 3
In this example, the Run file creates separate Graphics files containing p-charts for
each defect variable in the Data file NOZZLE.DAT. The Run file then creates a
Graphics file containing a Pareto diagram for four variables in the same Data file.
In the last portion of the Run file, the newly created Graphics files are used to
print charts with the PRINT command.
TITLE Print p-charts of nozzle defects
ATTRIBUT NOZZLE.DAT TDEFECT P G TDEFECT.NWG
ATTRIBUT NOZZLE.DAT SEAL P G SEAL.NWG
ATTRIBUT NOZZLE.DAT THREADS P G THREADS.NWG
ATTRIBUT NOZZLE.DAT WELD P G WELD.NWG
ATTRIBUT NOZZLE.DAT SURFACE P G SURFACE.NWG
PARETO NOZZLE.DAT M SEAL THREADS WELD SURFACE G NOZZLE.NWG
PRINT TDEFECT.NWG
PRINT SEAL.NWG
PRINT THREADS.NWG
PRINT WELD.NWG
PRINT SURFACE.NWG
PRINT NOZZLE.NWG
Note that this Run file could be broken into two Run files (between the PARETO
command and the first PRINT command), each of which could be executed at a
different time, perhaps by different users. This could be useful if, for example,
your data and your printer are not always available at the same time.
Example 4a
In this example, the Run file prompts the user for input using commands that
assign values to Replaceable Parameters. The user is prompted for the name of a
Data file, the name of a variable, and a date range. The Run file then creates a file
containing an Individual and a Range chart for the variable specified by the user
and displays it on the screen.
TITLE Generic IR with date filter
AskDATFile ^File1 "" "Select a Data file to use"
When this Run file executes, you will be prompted first for the name of the Data
file to use, then for the name of the variable to use, and then for the range of dates
to include in the analysis. If you specify ELEMENTS.DAT for the file name, HG
for the variable name, and 7/1/01 and 7/31/01, as shown below, the effect of the
example Run file will be the same as that of the following Run file:
TITLE Generic IR with date filter
IR ELEMENTS.DAT HG I R G $FILTER="DATE L 7/1/01 U 7/31/01"
CHART
Example 4b
This example illustrates the same task as the previous example. In this case,
however, the Run file prompts the user for input using commands that assign
values to “$” Replaceable Parameters.
TITLE Generic IR with date filter
PROMPT $FILE1 Enter the Data file to use
PROMPT $VAR1 Enter the variable to be charted
PROMPT $RANGEFILTER1 Enter the range of dates to analyze:
IR $FILE1 $VAR1 I R G $FILTER="DATE $RANGEFILTER1"
CHART
When this Run file executes, you will be prompted first for the name of the Data
file to use, then for the name of the variable to use, and then for the range of dates
to include in the analysis. If you specify ELEMENTS.DAT for the file name, HG
for the variable name, and 7/1/01 and 7/31/01, as shown below, the effect of the
example Run file will be the same as that of the following Run file:
TITLE Generic IR with date filter
Example 5
In this example, another software package runs the Quality Analyst Run File
Interpreter, invoking a Run file, and providing the file and variable names. A
second Run file is “nested” within the first, and its name is provided at run time.
The name of the nested Run file (and which additional charts are to be created) is
determined at run time by the originating software.
The following Run file setup will create an X-bar and Range chart and a Process
Capability chart for the variable ACID from the file TEST.DAT, start up a second
Run file to create a Run Chart for variables DSOLID and SSOLID, and return to
the originating Run file to create a report using all three variables in TEST.DAT.
Control returns to Windows (and hence, to the originating program) at the end of
Run file execution:
1 The originating program executes the following Windows command which
runs the Quality Analyst Run File Interpreter:
3 The nested Run file, BASIC2.RUN, which is specified in the command line
by the originating program, contains the following lines:
These two commands make sure that the settings you specify are properly written
to the Header file, regardless of the order in which you specify them.
Probably the best way to use these commands in a larger workflow is:
1 Use another program to create a temporary parameter file (the paramfile
argument described under CreateNWH and UpdateNWH).
2 In your Run file, use the “@” form of a CreateNWH and/or UpdateNWH
command and specify that temporary parameter file.
Select the Run file you want to process by clicking its name in the list, then
clicking Open.
Note: After you install Quality Analyst, if you install another program that re-
registers the “.RUN” file-name extension, you will not be able to execute
Quality Analyst Run files directly. See Registered File Types on page 14
for instructions on registering file-name extensions.
Simply create a command line that begins with the name of the Run file and
includes any additional command-line options.
For example, to execute the Run file CHART.RUN using data in the folder
E:\QA\DATA, the command line would be:
CHART.RUN $DATA="E:\QA\DATA"
Because Run files are associated with the Quality Analyst Run File Interpreter in
the Windows Registry, this command will launch the Interpreter and execute
CHART.RUN using the $DATA parameter shown.
However, if you subsequently install software that re-registers the RUN file-name
extension, such command lines will not work. In that case, you must use a
command line that includes the name of the Quality Analyst Run File Interpreter
(QARFI.EXE), a $RUN command, the name of a Run file, and any additional
command-line options.
For example, to execute the Run file CHART.RUN from another application or
from a program shortcut, the command line would be:
C:\QA\QARFI.EXE $RUN=CHART.RUN
See Command Line Options on page 502 for information about other command-
line parameters and Replaceable Parameters.
1 Right-click the Run file in Explorer, then select Create Shortcut from the pop-
up menu. (The shortcut usually appears at the end of the file list.)
2 [Optional] To relabel the shortcut, right-click it, select Rename from the pop-
up menu, enter a new label, then press <ENTER>.
3 Move the shortcut to a convenient location on your PC, such as the Windows
Desktop.
(See Passing Replaceable Parameters from the Command Line on page 486
for more information.)
Windows 98
1 In Windows Explorer, select Scheduled Tasks in the folder tree on the left.
2 In the list of tasks on the right, double-click Add Scheduled Task to start the
Scheduled Task Wizard.
3 When the list of applications appears, click Browse, locate QARFI.EXE in the
Quality Analyst program folder, and click OK.
7 Click OK.
Windows XP
1 Click Start, then select Control Panel.
2 Click the Scheduled Tasks icon(shown at left).
3 Double-click Add Scheduled Task to start the Scheduled Task Wizard.
4 On the first screen, click Next.
5 When the list of applications appears, click Browse, locate QARFI.EXE in the
Quality Analyst program folder, and click Open.
6 On the next screen, enter a name for the task. This will appear as the label of
the task icon in the Scheduled Task applet in Control Panel. Also specify the
execution frequency for the task, then click Next.
10 Click OK.
Syntax Explanation
$RUN=filename.run Name of Run file to execute.
$TITLE="text" System title text (replaces default title).
$REPORTDATE="text" Text to use for Report Date
$DATA=folder Folder (directory) containing Data files
$TEMP=folder Folder (directory) to store temporary files
$PLOT=folder Folder (directory) to store Graphics files
$QANWA=folder\QA.NWC Folder (directory) and file name of the Quality
Analyst configuration file
<Replaceable One or more Replaceable Parameters. See
Parameters> examples below.
@<command_file_name> Name of a file containing command line options.
(Used if options are too long for a command line.)
Note: Include quotation marks as shown. Also note that quotes around folder names and file names are
mandatory if any is long or contains a space.
For example, the following Windows command line will start the Quality Analyst
Run File Interpreter and execute the Run file ANALYZE.RUN. Every time the
Run File Interpreter encounters “^DatFileA,” “^VariableA,” or “^VariableB” in
ANALYZE.RUN, it will use the strings “ELEMENTS.DAT,” “PB,” and “CD”
respectively, instead.
QARFI $RUN=ANALYZE.RUN ^DatFileA="ELEMENTS.DAT"
^VariableA="PB" ^VariableB="CD"
Similarly, the following Windows command line will start the Quality Analyst
Run File Interpreter and execute the Run file ANALYZE.RUN. Every time the
Run File Interpreter encounters “^InFile,” “^Var1,” or “^Var2” in
ANALYZE.RUN, it will use the strings “ELEMENTS.DAT,” “PB,” and “CD”
respectively, instead.
QARFI $RUN=ANALYZE.RUN ^InFile="ELEMENTS.DAT" ^Var1="PB"
^Var2="CD"
command following an “@” character. For example, suppose you want to execute
the following Run file command.
QARFI $RUN=BASIC.RUN $TITLE="SEMIANNUAL REPORT: JANUARY-
JUNE 1996" ^DatFile1="JAN96.DAT"
^DatFile2="FEB96.DAT" ^DatFile3="MAR96.DAT"
^DatFile4="APR96.DAT" ^DatFile5="MAY96.DAT"
^DatFile6="JUN96.DAT" ^DateVar="DATE"
^ElemVar2="PB" ^ElemVar3="CD" ^ElemVar4="ZN"
^ElemVar5="HG" ^ElemVar6="CR" ^ElemVar7="MO"
^ElemVar8="MN" ^ElemVar9="AS"
Instead, you could create a file named SEMIANN.TXT with the following
contents:
$RUN=BASIC.RUN
$TITLE="SEMIANNUAL REPORT: JANUARY-JUNE 1996"
^DatFile1="JAN96.DAT"
^DatFile2="FEB96.DAT"
^DatFile3="MAR96.DAT"
^DatFile4="APR96.DAT"
^DatFile5="MAY96.DAT"
^DatFile6="JUN96.DAT"
^DateVar="DATE"
^ElemVar2="PB"
^ElemVar3="CD"
^ElemVar4="ZN"
^ElemVar5="HG"
^ElemVar6="CR"
^ElemVar7="MO"
^ElemVar8="MN"
^ElemVar9="AS"
the dialog box and resume the Run file. Clicking the Quit button will terminate the
execution of the Run file.
The dialog box shown below will appear at the completion of a CHART command
that does include the /DELAY option. One of the hands on the timer will move
each second. Click the Continue button to resume the Run file before the
programmed delay period expires. Click the Quit button to terminate the execution
of the Run file.
The dialog box shown below will appear when a PAUSE command is executed.
Click the Continue button to resume the Run file. Clicking Quit will terminate the
execution of the innermost Run file, if it is nested. Clicking Quit All will terminate
the execution of all Run files including any that are nested.
encountered depending on how Quality Analyst is configured (see Run File Error
Handling on page 248).
The following table contains example error messages, the error(s) they represent,
and explanations of the error and possible causes.
Tip: You can mitigate the effect of errors on the execution of lengthy Run
files by nesting Run files. See the discussion under RUN on page 467.
changes you should make, and issues you should be aware of, if you intend to use
Run files from earlier versions of Quality Analyst with newer versions.
This is particularly important if your Run files contain any commands that
perform file operations directly or indirectly, such as AskOpenFile, AskSaveFile,
COPY, DELETE, RENAME, or SHELL.
Note: See CreateNWH on page 433 and UpdateNWH on page 478 for special
considerations regarding the behavior of those commands on “.hed” and
“.nwh” files.
Any Graphics file created by a Run file command under Quality
Analyst 6.1 or later will contain XML-structured data, regardless of the
file-name extension (if any) used in the command. Thus, if you use an old
Run file with a command that creates a file named, for example,
PCAP.PLT, the content of that file will be in the new XML-structured
format of an “.nwg” file (even though the file-name extension is “.plt”),
and the Run file command will execute properly. However, even though
updating file names in such cases is optional, file management may
become confusing if you don’t.
to
XRS COOKIE.DAT WEIGHT X R G
CHART
CHAPTER 10
Importing Data
Quality Analyst lets you import data from other sources, such as other software or
data-acquisition systems. You can import data from:
comma-separated-value files (and similar files) in which the values that end
up in Quality Analyst’s columns are separated by commas in the source file.
Quality Analyst also “cleans up” files in which values are separated by
commas but which also contain other characters (like tabs, punctuation,
nested quotes, and so on).
fixed-field files in which the values that end up in Quality Analyst’s columns
are aligned in columns (using spaces for “padding”) in the source file.
Note: Some programs create space-delimited data files in which values are
separated by spaces and text values contain no spaces. Quality Analyst
can use these files without conversion. See Opening a Data File without
a Header File on page 518.
If you want to use Quality Analyst to work with data that are stored in the file
format of any of several popular database programs (including any ODBC-
compliant database), you can use the connectivity features of Quality Analyst to
connect directly to those databases without conversion. Database connectivity is
discussed in CHAPTER 12.
4/25/96,SBW,29.0,29.0
4/25/96,SBW,127.5,128.2
5/1/96,FS,172.0,175.0
5/9/96,SBW,55.5,55.0
5/9/96,FS,72.5,72.5
Comma-separated-value files, sometimes called comma-delimited files, often have
the file-name extension “CSV.”
When you import a CSV file, Quality Analyst performs the following tasks in this
order:
To import a CSV file, first close the Data Set in the Editor if one is open. Open the
File menu, select Import, then select CSV to Quality Analyst Format. The Convert
CSV to Quality Analyst Format dialog box will appear.
In the area labeled Input File, click Select File to select the external file to be
converted. Its name will then appear in the box.
In the area labeled Output File, click Select File to specify a name and location for
the converted file. Its name will then appear in the box.
In the box labeled Additional characters to remove, enter any characters you want
removed from the file during conversion, in addition to those described in the
steps above. Enter the characters one right after the other; do not separate them
with spaces, commas, or any other character. Each character you enter will be
removed individually (not as a group) from the input file. For example, if your
input file contains values enclosed in brackets, enter the two characters [ and ].
Note: You must not include a comma or space here since Quality Analyst
handles those specially.
In the area labeled Columns to check, specify how many columns of data are in the
source file. This setting affects only the data-conversion report that Quality
Analyst displays after the conversion. Regardless of which option you select here,
you will end up with the same converted file. The data-conversion report will list
any discrepancies between the number of fields expected and the number of fields
actually found in the file.
Check Columns to specify the number of data columns Quality Analyst
should “look for” when converting the input file. Enter the number of
columns in the adjacent box.
Check Use first row to use the number of data elements found in the first row
in the input file as the number of columns to “look for.”
Check No test to forgo a check on the number of columns. If you select this
option, no discrepancies in column count will be reported.
Click OK to convert the file. The import function does not actually load the file
into the Editor. After converting the file, you must open it as you would any other
Quality Analyst Data Set. See Opening a Data File without a Header File on
page 518.
4/24/96FS 152.0153.0
4/25/96SBW 29.0 29.0
4/25/96SBW127.5128.2
5/1/96FS 172.0175.0
5/9/96SBW 55.5 55.0
5/9/96FS 72.5 72.5
To import a fixed-field file, first close the Data Set in the Editor, if one is open.
Open the File menu, select Import, then select Fixed Field to Quality Analyst Format.
The Fixed-Field to Quality Analyst Format dialog box will appear.
In the area labeled Input File, click Select File to select the external file to be
converted. Its name will then appear in the box.
In the area labeled Output File, click Select File to specify a name and location for
the converted file. Its name will then appear in the box.
Quality Analyst uses a “Fixed-Field Definition” to determine which columns of
characters in the fixed-field file represent data (field) columns. Fixed-field
definitions can be saved in files (with FFD extensions) and reused later (for
example, in Quality Analyst Run files). In the area labeled Fixed Field Definition
File, you can create and edit Fixed-Field Definitions and optionally save and open
them.
Click Select to select a Fixed-Field Definition (FFD) file that you have
already created. Its name will appear in the box.
Click Edit to modify the definition contained in the file you have selected.
(See instructions below.)
Click Create to create a new Fixed-Field Definition. (See instructions below.)
Click Save to save the current Fixed-Field Definition in the file whose name
is displayed in the box.
Click Save As to save the current Fixed-Field Definition in a file other than
the one named in the box.
If you click OK without saving the Fixed-Field Definition to a file, a message box
will appear asking if you want to save the Fixed-Field Definition. You need not
save a Fixed-Field Definition to an FFD file, but doing so can be useful if you:
repeatedly receive fixed-field files to convert that have the same column
layout, or
want to convert fixed-field files in a Run file (in which case an existing FFD
file is required).
During the conversion, any spaces within a field are converted to underscores
(“_”). When the file has been converted, Quality Analyst will display a data-
conversion report that contains a summary of actions taken and lists any
discrepancies found during the conversion.
The import routine does not actually load the file into the Editor. After converting
the file, you must open it as you would any other Quality Analyst Data Set. See
Opening a Data File without a Header File on page 518.
Use this form to specify the column locations in a fixed-field file for importing
into Quality Analyst format.
Specifying Columns
For each formatted column in the input file, specify:
the Starting Character Position of the column (the first position is 1).
the Width of the column, in characters.
Quality Analyst will recalculate the starting positions based on the widths.
Though the columns you define must not overlap, they need not abut. As in the
example below, you can define a “null” area between columns. Quality Analyst
will ignore anything found in the input file that is not within a defined column.
This is useful if you want to ignore one or more data columns or label columns in
the input file.
This dialog box displays all Data files. Select the Data file you want to open. The
No Data Set Structure dialog box will appear.
This dialog box lists all Header files. Select the Header file that contains the
structure you want to apply to the Data file. When you click Open, the Editor will
display the data in the Data file with the data structure from the Header file.
If you inadvertently select a Header file that Quality Analyst cannot apply to the
Data file, a warning message will appear. Click OK and select another Header file.
Quality Analyst will assign temporary variable names and the alphabetic data type
(“A”) to each column. Simply redefine the structure of the Data Set as explained
under Column Definition and Calculations on page 72.
Tip: Because Quality Analyst also loads the data into the Editor, you can
reposition the windows as shown in the figure to view the data as you
restructure it.
2 Under Input File, click Select File and specify the Quality Analyst file to
convert.
3 Under Output File, click Select File and specify a name and location for the
destination comma-delimited file.
4 Check Delimit Text with Quotes if you want to enclose non-numeric fields in
quotation marks.
5 Check Variable Names on first line if you want to write the variable names as
the first line of the file.
6 Click OK.
menu and select either Variable Definition and Specifications or Column Definition
and Calculations.)
The cells that you paste into may be empty or may contain either data or the
missing-data symbol. Existing data and the missing-data symbol will be
overwritten by pasted data. If Quality Analyst adds new rows when you paste, any
new cells that do not receive pasted data will contain the missing-data symbol.
The following example explains pasting data into the Editor. We will paste eight
rows and four columns of data to a Quality Analyst file that already has dates in
the first column.
Editor with eight rows and four columns free for importing
In the source application (in this example, Microsoft Excel) we mark and copy the
desired data.
In Quality Analyst, first position the cell in the upper left corner of the range you
want to paste to, then open the Edit menu and select Paste (or press <CTRL+V>.
The data will appear in the range you cleared for them and the Verify Data message
box will appear.
Select Yes to have Quality Analyst verify the validity of the imported data. If you
select No, we strongly recommend that you verify the pasted data before
performing any analysis or charting with them. (See Verifying Data on page 100.)
Open the Edit menu and select Copy to copy the highlighted data to the clipboard.
Or open the Edit menu and select Cut to copy the data to the clipboard and remove
it from the Editor.
In the destination application, follow its instructions for pasting data. Usually, this
involves selecting the cell in the upper left corner of the destination range, then
opening the Edit menu and selecting Paste, just as we did when pasting data into
the Editor. The following example uses Microsoft Excel.
CHAPTER 11
If you had a Data Set open in the Editor when you started this procedure, the box
in the area labeled File to Modify will display the name of the Data file. If not, you
should click Select File and select the file whose missing-data symbol you want to
change.
The box labeled Current missing data symbol displays the current System
Parameter for the missing-data symbol (see Missing Data Symbol on page 272).
This might not be the same as the missing-data symbol in the file you want to
modify.
In the box labeled Old Missing Data Symbol, enter the symbol or text string
currently used to indicate missing data in the Data Set you want to modify.
In the box labeled New Missing Data Symbol, enter the desired symbol or text
string to indicate missing data. The missing-data symbol can be a single character
(such as “*”) or a string of text (such as “N/A” or “MISSING”). Spaces and
commas are not allowed within the missing-data symbol.
To rotate a Data Set, open the File menu, select Utilities, then select Rotate Data
File. The Rotate Data File dialog box will appear.
If you had a Data Set open in the Editor when you started this procedure, the box
in the area labeled Input File will display the name of that Data Set. If not, you
should click Select File and select the file you want to rotate.
In the area labeled Output File, click Select File to specify the name of the file to
create from the new arrangement of rows and columns.
In the area labeled Rotation, specify the rows or columns you want to rotate.
Select All Columns to include all columns (and rows) in the rotation.
Select Select Columns to include only those columns that you specify (and all
rows) in the rotation. In the adjacent box, enter the column numbers,
separated by spaces, to include in the rotation. For example, 1 2 3 4
Select All Rows to include all rows (and all columns) in the rotation.
Select Select Rows to include only the rows you specify (and all columns) in
the rotation. In the adjacent box, enter the row numbers (separated by spaces)
of the rows you want to include. For example: 1 3 4 5 7.
Notes: When you try to open the new rotated file, Quality Analyst will tell you
that the data has no structure (that is, there is no Header file for the Data
file). At that time, you can either create a new structure or use the
structure of another Data Set, as if you were creating a brand new
Data Set.
In both cases, Quality Analyst retains the original file and creates a new file
containing the converted data.
In the area labeled Conversion Type, select Convert from single column. In the
adjacent box, enter the Number of observations per sample. This is the number of
data columns (excluding description columns) you want in the new multi-column
Data Set.
Check Include description columns if you want to use the values of one or more
fields in the existing single-column Data Set as description field(s) in the new
multi-column Data Set. (In the input file, if the value of the description field varies
within each set of rows that will be converted to columns, the value from the first
row of the set will be used to “label” the row in the new multi-column Data Set.
See diagram above.)
When you click OK, the Convert Subgroups to/from Single Column variable-
selection dialog box will appear. (The lower half of this dialog box, labeled
Description Columns, will appear only if you selected Include description columns
above.)
Select the Column to Convert (as explained under Selecting a Single Variable on
page 121) and, if applicable, the Description Columns to use (as explained under
Selecting Multiple Variables on page 123).
The new multi-column Data Set will include a Header file and the columns will be
named VAR_1, VAR_2, and so on. You can open the file with the Editor and
restructure it as needed.
Notes: In the input file, if the value of the description field varies within each set
of rows that will be converted to columns, the value from the first row of
the set will be used to “label” the row in the new multi-column Data Set.
Select the Columns to Convert and, if applicable, the Description Columns to use.
(See Selecting Multiple Variables on page 123).
Click OK to convert the file. The converted column in the new Data Set will be
named “VAR_N” where “N” is one more than the number of description columns
you selected. You can open the new Data Set with the Editor and restructure it as
needed.
Note: In fact, the primary purpose of this feature of the Quality Analyst Editor
is to create and maintain SEL files for use by the EXTRACT command in
Run files. It is probably easier to eliminate unneeded or unwanted rows
of data by saving a filtered Data Set, as explained under Filtering Data
starting on page 90.
Extracting Data
To extract part of a Data Set, first close the Data Set in the Editor if one is open.
Open the File menu, select Utilities, and select Extract Subset from Data File. The
Extract Subset from Data File dialog box will appear.
In the area labeled Input File, click Select File to select an input file. This is the
existing file from which you want to extract data.
In the area labeled Output File, click Select File to specify a name and location for
the new file that will contain the extracted data.
In the area labeled Extract Definition File, click Select to load extraction criteria
from an existing Extract Definition (SEL) file. The name of the file will be
displayed in the text box. Click Edit to modify the extraction criteria. Click Create
to create a new set of extraction criteria. Click Save to save the current extraction
criteria in the current Extract Definition file. Click Save As to save the current
extraction criteria in a new file. (For instructions on what to do after you click
Create or Edit, see Creating and Editing an Extract Definition File below.)
To extract the subset, click OK. The new extracted Data Set will have a Header file
derived from the Header file of the parent Data Set—the same file title and
column definitions will be used.
Use this dialog box to define criteria for extracting a subset of a Data Set. You can
define up to ten criteria.
The Definition List contains the current set of definitions. They will be named
Definition 1, Definition 2, and so on. As you select definitions in this list, the label
of the area to the right will change to the name of the selected definition and the
components of the criterion will be displayed.
To add a new definition to the end of the list, click Add. To delete an existing
definition, select it in the Definition List, then click Delete. To insert a new
definition in a specific position in the list, select the definition above which you
want to insert the new one, then click Insert.
In the area labeled Definition N, specify the extraction criterion for the definition
selected in the list to the left.
Select And if the extracted data must meet both this definition and the previous
definition. Select Or if the extracted data must meet either this definition or the
previous definition. (This option is available for all but the first definition.)
From the drop-down list labeled Variable, select the data column on which you
want to base this extraction criterion.
Select Match to extract only those values that exactly equal the value you enter in
the adjacent box. Select From to specify a range of values to extract. Enter a lower
limit for the values you want to extract in the adjacent box. Place a “>” before the
value to exclude records where the field value is exactly equal to the lower limit.
Values “just greater than” the lower limit will be included. In the box labeled To,
specify an upper limit for the values you want to extract. Place a “<” before the
value to exclude records where the field value is exactly equal to the upper limit.
Values “just less than” the upper limit will be included.
Note: To specify an open-ended range of values, specify values for either From
or To, but not both. For example, you could specify values greater than or
equal to 25 or less than 0.05:
Select Positive to extract data that meet the criterion you have defined. Select
Negative to extract all data except those that meet the criterion you have defined.
When you click OK, Quality Analyst will validate all definitions and identify any
illegal or missing components for you to correct.
Note: Adding a calculated variable to a file is much easier using the procedure
described under Defining Calculated Variables on page 62.
You can probably skip this section unless you use existing Run files that
contain many NEWVAR commands, and those files are too complex for
you to consider modifying them.
In the area labeled Input File, click Select File to select the existing Data Set that
contains the variable(s) from which you want to calculate one or more new
variables. The name of the file will be displayed in the text box.
Output to is discussed below.
In the area labeled Variable Definition File, click Select to load variable definitions
from an existing Variable Definition (VDF) file. The name of the file will be
displayed in the text box. Click Edit to modify the variable definitions. Click
Create to create a new set of variable definitions. Click Save to save the current
variable definitions in the current Variable Definition file. Click Save As to save
the current variable definitions in a new file.
After you have saved the Variable Definition file, you can click Cancel if you are
using this routine only to create a Variable Definition file for use in a Run file. Or,
see Saving New Calculated Variables to a Data File below.
This dialog box looks and acts just like the Column Definition and Calculations
dialog box described under Specifying Column Definition and Calculations on
page 61. Follow those instructions and guidelines as you compose formulas for the
new variables in this dialog box.
CHAPTER 12
Database Connectivity
Overview
NWA Quality Analyst lets you define Quality Analyst Data Sets that use external
databases as the source of data. This lets you connect to external data—either
directly or via ODBC (Open DataBase Connectivity)—while retaining the
complete functionality of Quality Analyst. This version of Quality Analyst
provides seamless connections to any ODBC-compliant database and direct
connections to many PC-based databases.
Using Quality Analyst, you define a connection between a Quality Analyst Data
Set and an external data source. Once the link is established, using the external
data is transparent to you. Quality Analyst stores the information needed to
establish the connection to the external data source as part of the Data Set
definition. Whenever you use a Connectivity Data Set, Quality Analyst simply
uses the definition to read the data from the external data source instead of from a
Quality Analyst Data file. Quality Analyst Connectivity Data Sets are listed
alongside the standard Quality Analyst Data Sets in the Open dialog box.
You can also configure Quality Analyst to read the specifications for the variables
in your Data Set from an external database. When you do so, the specifications are
read whenever you open the Data Set, and whenever a CONNECT command
(page 432) is executed in a Run file. See Reading Specifications from a Database
on page 46 for instructions.
Note: Refer to the Installation Guide for important information about installing
and configuring the connectivity components of NWA Quality Analyst.
Direct Connection
Quality Analyst can connect directly to “desktop” databases created by dBase,
FoxPro, Microsoft Access, and Paradox. It can also connect directly to data stored
in Microsoft Excel spreadsheets.
The Quality Analyst installation program installs the Microsoft Data Access
Components (MDAC) which handles all direct connections to desktop databases.
Installing additional drivers is not necessary.
ODBC Connection
Quality Analyst should connect to virtually any data source that supports ODBC.
To connect to ODBC-compliant databases, you must:
Have the appropriate ODBC drivers installed on your system
Define an ODBC data source for each database you want to read (as
discussed below)
Use Quality Analyst to create the definition. A convenient user interface lets
you select each component from lists provided by querying the data source.
You may select fields only from a single table in the data source selected.
You can define any number of filters to limit the amount of data retrieved
from the data source or to sort it on any field as it is retrieved.
Provide a complete SQL statement that performs all the functions necessary
to retrieve the desired data from the database. You can enter the SQL
statement by composing it at the keyboard or by copying and pasting it from
other applications. You should use this method if you need to perform
multitable operations or complex manipulation to retrieve the correct data.
(The Quality Analyst connectivity filters can not be used with the SQL
statement.) Although Quality Analyst performs some syntax checking, you
are responsible for the SQL statement and the consequences of its use.
Defining a Connectivity Data Set with Quality Analyst involves these steps:
1 Select New External Database when creating a new Data Set.
2 Specify a connection type (either ODBC, or a specific database for direct
connection).
3 Specify an external data source.
4 Do one of the following:
Select a table from the external data source
or
Enter SQL commands to define the data within the data source to use.
5 Assign fields from the data source to columns in the Quality Analyst Data Set.
6 Identify the key columns in the Data Set.
7 Specify data filters to limit the amount of data retrieved. (This step is
available only if you select a table in step 4, and is optional.)
8 Specify how the data should be sorted as it is retrieved. (This step is available
only if you select a table in step 4, and is optional.)
Select New External Database. The Quality Analyst External Database Setup dialog
box will appear.
The dialog box lists files of the selected database type (*.MDB for Microsoft
Access, for example). Select the external Data file you want to connect to. Make
sure it is a file created by the database application you selected in the previous
step. After you click OK, the name of the file you select will be displayed in the
box labeled Data Source.
ODBC Sources
If you selected an ODBC connection, select the source of the data you want to use
for the Data Set from the drop-down list. This list will contain all ODBC data
sources available on your system. If no sources are listed, or if the desired source
is not listed, you need to create an ODBC source definition using the ODBC
module of Windows Control Panel. See ODBC Connection on page 543.
When you select a table, the names of the fields in that table will be displayed in
the area of the dialog box labeled Available Fields. You can use this list of fields to
confirm that you have selected the correct table.
(If you do not need to denormalize the data table, skip to Setting External
Database Access Options on page 557.)
Overview of Denormalization
Many data-collection systems store data in “normalized” databases where each
database record includes a data identifier and a measured value. Thus, each
reading occupies several records in the database, as illustrated here:
“Denormalizing” this data turns each multi-record reading into a single record
with a column (field) for each unique value of the data identifier. For example,
denormalizing the data above would produce this data:
Notice that each unique Date-SerNum pair identifies a separate reading, each
unique value of ID becomes a separate data column, and the original Values for
each ID become the data in the columns. This arrangement makes the data suitable
for use by Quality Analyst.
Quality Analyst can denormalize numerical data (as shown in the example above)
as well as textual data, as shown in the following examples.
In the data table below, each row contains a single observation of a particular
defect on a particular day.
Date Defect
1/24/2005 Bent
1/24/2005 Scratched
1/24/2005 Bent
1/24/2005 Bent
1/24/2005 Scratched
1/25/2005 Dirty
1/25/2005 Scratched
1/25/2005 Scratched
1/25/2005 Dirty
1/25/2005 Scratched
1/25/2005 Bent
1/25/2005 Scratched
1/25/2005 Scratched
1/26/2005 Scratched
1/26/2005 Dirty
1/26/2005 Dirty
1/26/2005 Bent
1/26/2005 Torn
1/26/2005 Scratched
1/26/2005 Scratched
1/26/2005 Scratched
1/26/2005 Scratched
1/26/2005 Scratched
1/26/2005 Dirty
Quality Analyst can denormalize this data such that each row is a single date, each
type of defect is a separate column, and the data is the number of occurrences of
that defect on that day. Denormalizing the data above would produce this data:
Quality Analyst can denormalize this data such that each row contains a single
date, each shift is a separate column, and the data is the goal status for each shift
on that day. Denormalizing the data above would produce this data:
In the Data Field list, select the field in the database that contains the measured
data—that is, the Value , Defect, and Goal Status columns in the examples
above.
In the Column Names Field list, select the field in the database that contains the
data identifiers—that is, the ID , Defect, and Shift columns in the examples
above. (Yes, in the second example, the Data Field and the Column Names Field are
the same.)
unique values to display. To display all values, select All. To display only the most
frequently recurring values, select Top, then enter the number of unique values to
display.
If you are connecting to a very large database, be aware that querying the database
to determine those unique values may take more time than you want. So before
you display all column names, you may want to filter the database to limit the data
retrieved to records whose field values match certain criteria. To do this, click the
Filter Names button. The Filter Records to Search for Names dialog box will
appear.
In the drop-down list, select the variable on whose values you want to base your
filter.
3 In the adjacent box, enter the lower limit of the desired range.
4 Next to To:, select an upper-limit condition from drop-down list:
Select “=” to include records where the field value is exactly equal to the
upper limit.
Select “<” to exclude records where the field value is exactly equal to the
upper limit. Values “just less than” the upper limit will be included.
5 In the adjacent box, enter the upper limit of the desired range.
Click OK.
(Note: If you change the List Frequent Names setting after clicking Display
Names, you must click Display Names again to update the list.)
In the Column Names list, select the check boxes next to the values that you want
to turn into data columns in the Data Set. To select all displayed values, click the
Select All Names button.
taken together, is unique. For example, you could configure Quality Analyst to
group records for each unique combination of a date value, a time value, and a lot
number.
In the Group Rows by area, click the New button to select the first grouping field.
The Select Grouping Field dialog box will appear.
This dialog box displays all allowable fields. For date and time fields, there are
additional listings with a grouping parameter, for example month or hour. If you
select one of these “derived” fields, Quality Analyst will not use the exact value of
the field to group records; instead it will group records according to the
supplemental parameter. For example, if you select [Time] (hour), then all database
records from the same hour will become a single record in the Data Set. When you
group by (week), values from each week (Sunday through Saturday) will be
grouped into a single record.
Select the field you want to use, then click OK. Use the New button again to
specify additional grouping fields. To remove a grouping field, select it in the
Group Rows by list, then click Remove.
If the grouping fields you select cause more than one database value to be grouped
into a Data Set record, the value that Quality Analyst uses for the grouped record
depends on the data type of the Data Field:
Numerical Data Quality Analyst computes the mean of all grouped
values.
Text Data Quality Analyst uses the setting in the area labeled Report
Results As (which will appear whenever the Data Field is a text data type).
Select Count if you want Quality Analyst to count the number occurrences of
unique values in that field, as in the Defect example above.
Select Text if you want Quality Analyst to simply return the text value of the
field, as in the Goal Status example above. Note: Because it is not
meaningful to “group” multiple textual values, select Text only if your
settings will ensure that only a single value will be retrieved during grouping.
Otherwise, the denormalized data will not be meaningful.
When you have specified denormalization settings, click OK and skip to Setting
External Database Access Options on page 557.
Enter the SQL commands that perform the operations required to define the data
to which you will link your Data Set. (You can use the Paste button on the right to
insert SQL statements copied to the Windows Clipboard from other applications.)
Tip: If you type the statements into the dialog box directly, begin a new line
by pressing <CTRL+ENTER>.
Note: You can use Quality Analyst’s Replaceable Parameters in your SQL
statements for added flexibility in querying the source data. See Using
Replaceable Parameters with Filter Criteria in SQL Statements starting
on page 573 for details.
To select a table, highlight it in the list of Database Tables on the left, and click
Select. The name of the table will move to the list of Selected Tables on the right.
Refer to the section Configuring and Troubleshooting, beginning on page 587,
for more information.
Select System Default to use the global settings. (Configuring these global settings
is explained under Configuring Connectivity on page 587.)
Select Custom to override the global parameters with the parameters you specify
in this dialog box. The parameters will be saved for the current Data Set in its
Header file.
For explanations of these settings, see Data Retrieval on page 590 and Database
Options on page 588.
You define your new Data Set by specifying a name and data type for each
column in the Data Set and then assigning a field from the external data source to
the data column using the External Database Fields tab.
Select the External Database Fields tab at the bottom of the dialog box. At the top
of the tab, the Connection type, Data Source, and Table are displayed. Below that,
all Fields requested from the data source are listed in the format
[table].[field_name]. You need not include all of these fields in your Data Set.
You can fill out the Database Connectivity Definition form in any order you find
convenient. For example, you can:
select all fields at once (using the Assign All button), modify the automatically
assigned variable names and data types, if desired, and add descriptions; or
enter all the column names, data types, and descriptions, then assign database
fields; or
select all applicable database fields first, then enter a name, data type, and
description for each; or
fill in one row at a time, entering a name and data type, assigning a database
field, and entering a description for each as you go.
The field name will appear as the External Database Field in the Database
Connectivity Definition form.
In the following figure, the fields “DATE” and “TIME” have been selected for the
first two data columns and the field “LOTCODE” is about to be selected for the
third.
After you have assigned External Database fields to Data Set columns, enter a
variable name and data type for each variable.
The variable name you enter must conform to Quality Analyst rules for variable
names:
They must be 2 to 32 characters long.
They may contain letters (“A”–“Z”), numerals (“0”–“9”), or the underscore
(“_”).
They must begin with a letter.
Each must be unique within a Data file.
The data type must be a legal Quality Analyst data type. (See Data Type on
page 37.) External database field names, on the other hand, need not conform to
Quality Analyst’s rules for naming variables.
You can then edit the column names (if you want) and specify data types for each
data column. (Assign All will overwrite any existing contents in the column
definition form.)
Note: Using Assign All can result in duplicate field names if, for example, the
first 32 characters of two external database fields with longer names are
the same. When you click OK, Quality Analyst will detect this and ask
you to correct it.
10
If you want your DateTime variable (type D) to contain only date data without the time
portion, you must first change a system setting as explained in Handling Date-and-Time
Fields on page 590.
If you want to store time in a separate Data Set column, Quality Analyst can
extract the time portion of a DateTime database field and write it to a time variable
(type T).
(By default, Quality Analyst will retrieve the entire table from the data source
when you open the Data Set.) The filter criteria you specify here become part of
the data definition stored in the Header file of the Data Set. When you open an
external Data Set that includes filter criteria, Quality Analyst translates the criteria
into SQL commands and sends them to the external source so the data are filtered
by the external source before being sent back to Quality Analyst.
Note: The External Database Filter tab is only accessible if your Data Set is
based on the single table you selected from the drop-down list in the
External Database Setup dialog box. If you chose instead to base your
Data Set on an SQL query, the External Database Filters tab will not be
available and the SQL commands you enter must perform all necessary
filtering.
The field name will appear in the list of Filters on the left and in the Column
box.
2 Select Match.
3 Do one of the following:
Enter a value to match in the adjacent box. This must be a legal value for
that variable type.
Select (Ask) from the drop-down list if you want to specify this filter
criterion each time you use the database. (See Using the (Ask) Settings
on page 569.)
Here is an example of a match filter that will retrieve only data produced on
July 1, 2001.
In this example, the user will be prompted for a match date when the database is
opened.
Here is an example of a filter that will retrieve only data records in which the
LOTCODE field begins with any three characters and ends with -27-F.
Here is an example of a range filter that will retrieve only data collected during
July 2001.
You can define an “open-ended” range filter by leaving either the To or the From
entry blank. Here is an example of a range filter that will retrieve only data
collected before August 1, 2001 (“To <”).
Here is an example of a range filter that will retrieve only data collected on or
after June 1, 2001 (“From =”).
When you connect to the database, either by opening the Data Set with the Quality
Analyst Editor or by using a CONNECT command in a Run file, you will be
prompted for the appropriate filter criterion in an appropriately labeled dialog box.
(The examples under Prompting for Filter Criteria at Run Time on page 574
show samples of these dialog boxes.)
If you do not want to be prompted for this filter criterion, but would rather pass
values from the command line or set values within a Run file, you must use this
same syntax in the command line or the SET command. For example:
YES: SET ^DATE!MatchFilter = "10/1/2005"
NO: SET ^DATE = "10/1/2005"
NO: SET ^BeginDate="10/1/2005"
YES: QARFI.EXE $RUN=WTTASK.RUN
^DATE!MatchFilter="10/1/2005"
NO: QARFI.EXE $RUN=WTTASK.RUN ^DATE="10/1/2005"
NO: QARFI.EXE $RUN=WTTASK.RUN ^BeginDate="10/1/2005"
Removing Filters
To remove a connectivity filter:
1 Select the filter in the list of Filters at the far left of the dialog box.
2 Click Remove below the list.
Quality Analyst uses the key columns to sort the data after retrieving it from the
database, so you must arrange the key columns in order of decreasing sorting
significance—that is, most significant (coarse) first, and least significant (fine)
last.
The things that Quality Analyst uses key columns for are listed under Specifying
Key Columns for the Data Set on page 57. We strongly recommend that you
identify key columns for your Data Sets in all cases, but if you are certain you will
never use any of the features listed on page 57, you can forego selecting any key
columns. Simply select the No Key Columns check box before clicking OK in the
Database Connectivity Definition dialog box.
Note: You can specify key columns either in the Database Connectivity
Definition dialog box, or on the Key Columns tab of the Variable Definition
and Specifications dialog box.
Quality Analyst will save the connectivity information in the Header file of the
new Connectivity Data Set. When you close the Data Set, the Data file will be
retained on disk. Each time Quality Analyst uses the Data Set, it will try to reread
the data from the external source. If it is unable to do so, it will offer to use the
Data file retained from the last successful retrieval.
Note: Now that you have connected your Data Set to your database, you should
configure the other parameters of the Data Set. See Specifying Variable
Definition and Specifications on page 35 to specify variable
descriptions, specifications, Input Masks, numeric formats, axis labels,
and assignable causes and corrective actions. See Specifying Column
Definition and Calculations on page 61 if you need to define additional
calculated variables.
Note: This example statement shows the Microsoft Access SQL syntax for
denoting a date constant (it is enclosed by two “#” characters) which
differs from the syntax of other varieties of SQL. If you are passing the
query to an external database (see Opening Databases on page 588) that
uses a different variety of SQL, you must use the correct syntax for that
variety of SQL in all your statements. Consult your database’s
documentation for more information.
This statement retrieves the fields PRODUCT and WT from a database named DB
for records where the value of the DATE field is on or later than 10/1/2004.
But rather than “hard-wire” the date 10/1/2004 into the query, Quality Analyst lets
you supply the date each time the database is queried. There are three ways you
can supply this value:
by having Quality Analyst prompt for it at run time,
by setting it within a Run file, or
by setting it on the command line for the Run File Interpreter.
range. ‡
“!LastDaysFilter” prompts for the number of
recent days to include.
“!LastMonthsFilter” prompts for the number of
recent months to include.
“!LikeFilter” prompts for a “mask” or “template”
with wildcards.
“!MatchFilter” prompts for an exact value for the
variable. ‡
“!ToFilter” prompts for the ending value of a range.
‡
† If the column is one of a multicolumn subgroup, add a colon and the column number to the variable
name (for example, WEIGHT:3).
‡ Wording in the prompt will reflect the data type of the variable and whether the value entered will be
included or excluded.
(see note on SQL syntax on page 574) and you start the Run File Interpreter with
the command line:
QARFI.EXE $RUN=LUMBTASK.RUN
Note that the dialog box displays the variable name (Date), the description of the
variable (“Date Produced”) taken from the Header file of the Data Set, and
appropriate labeling for the data-entry box based on the data type (DateTime) and
the “>=” preceding the first “#” in the SQL statement. If the user enters 6/1/04
and clicks OK, Quality Analyst replaces “^Date!FromFilter” in the SQL
statement with “6/1/04” and passes the SQL statement to the database as:
SELECT * FROM lumber WHERE Date>=#6/1/04#
Here are some examples of other prompting dialog boxes and the SQL statements
that produce them. (See note on SQL syntax on page 574.)
This SQL statement:
SELECT * FROM Lumber WHERE Date<#^Date!ToFilter#
(see note on SQL syntax on page 574) and you start the Run File Interpreter with
the command line:
QARFI.EXE $RUN=WTTASK.RUN
then Quality Analyst will take “10/1/99” from the SET command, substitute it
for “^BegDate” in the SQL statement, and pass the SQL statement to the
database as:
SELECT PRODUCT, WT FROM DB WHERE DATE>=#10/1/99#
CONNECT WEIGHT.DAT
(see note on SQL syntax on page 574) and you start the Run File Interpreter with
the command line:
QARFI.EXE $RUN=WTTASK.RUN ^FromDate ="10/1/04"
then Quality Analyst will take “10/1/04” from the command line and substitute
it for the Replaceable Parameter “^FromDate” in the SQL statement and pass the
SQL statement to the database as:
SELECT PRODUCT, WT FROM DB WHERE DATE>=#10/1/04#
the Data Set to prompt you for any filter criteria (by specifying (Ask) as a filter
value, for example), one (or more) of several possible dialog boxes may appear.
(See Prompting for Filter Criteria at Run Time on page 574 for examples of
possible dialog boxes.)
After the data is retrieved from the source, Quality Analyst stores it as a standard
Quality Analyst Data Set, so you will experience the same performance level, after
the file is open, as when you work with a standard Quality Analyst Data Set.
After opening a Connectivity Data Set, you may want or need to update the local
data periodically by re-querying the database. To do this, open the Data menu and
select Query Database.
Also, when Quality Analyst opens a Connectivity Data Set, it will use the
connection definition stored in the Header file of the Data Set. If the structure or
location of the database has changed since the Data Set was created, Quality
Analyst may have problems retrieving the data. The error messages produced may
not be useful in identifying the source of the error since the problem may be the
database application, a network problem, configuration changes, and so on. If you
encounter such problems, refer to Configuring and Troubleshooting beginning
on page 587, or contact your Information Systems administrative or support staff
before calling NWA.
As you change the data definition, you may experience delays or pop-up
messages. This is because Quality Analyst may connect and reconnect to the data
source, sending commands and receiving results, as you redefine the Data Set. The
nature and length of these delays will depend on the type and location of the data
source. Remember, though, that once your Data Set is defined and opened, you
can expect the same performance from a Connectivity Data Set as from a standard
Quality Analyst Data Set.
To change the External Database definition, open the Edit menu and select
Database Connectivity Definition. The Database Connectivity Definition dialog box
will appear.
It should look just as it did when you finished defining the External Database
columns when you created the Data Set. Remember, the rows in this form define
columns in the Data Set.
To change the Connection, Data Source, or Table for the Data Set:
1 Click the External button. The External Database Setup dialog box will appear.
Note: If you change the Data Source or Table, all field references and data
filters will be deleted. The Quality Analyst column names will be
retained but you may need to modify them to match the fields you
choose from the new source or table for the new data definition.
Note: Be sure to redefine any filters that were based on fields you
removed.
When you are done redefining the Data Set, select OK. The data from the previous
definition will be discarded and the new connectivity definition will be used to
create a new Data file.
Note: If the CONNECT command is not used, the data present in each Data Set
at the time the Run file is executed will be used. This will likely be “left
over” from the last time the Data Set was used.
The CONNECT command for a Data Set may occur anywhere in the Run file as
long as it is before the Data Set is first referenced for processing.
Tip: Put all CONNECT commands at the top of a Run file where they are
easily located and sure to be processed before the Data Sets are used.
To avoid being prompted for input at run time, you can include SET commands in
your Run file to “pre-assign” values to the Replaceable Parameters in the SQL.
Pre-assigning filter criteria is illustrated under Setting Filter Criteria within Run
files on page 578. But you can also pre-assign other Replaceable Parameters, as
illustrated in the last example below.
Read more about Replaceable Parameters starting on page 484.
Examples
In the following Run file, the data for the Data Set named DIAMETER is
downloaded before it is used to create an IR chart:
CONNECT DIAMETER.DAT
IR DIAMETER.DAT DIAM I R G
CHART
In the following Run file, the data for the Data Set named PLATES, which
contains calculated variables, is downloaded and calculated before it is used to
create an IR chart:
CONNECT PLATES.DAT
CALC PLATES.DAT
IR PLATES.DAT THICK I R G
CHART
In the following Run file, four different Data Sets are used. Three of them
(LINE1.DAT, LINE2.DAT, and LAB.DAT) are Connectivity Data Sets and one
(TEST.DAT) is a standard Quality Analyst Data Set. Note that the CONNECT
command for LAB.DAT could have appeared at the top of the file:
CONNECT LINE1.DAT
CONNECT LINE2.DAT
XRS LINE1.DAT WEIGHT X R G
CHART
XRS LINE2.DAT WEIGHT X R G
CHART
IR TEST.DAT PH I R G ONE.NWG
PCAP TEST.DAT PH G TWO.NWG
CONNECT LAB.DAT
IR LAB.DAT VISC I R G THREE.NWG
PCAP LAB.DAT VISC G FOUR.NWG
CHART GROUP 2V2H ONE.NWG TWO.NWG THREE.NWG FOUR.NWG
Let’s say you have a Run file named WTTASK.RUN that connects to a
Connectivity Data Set named WEIGHT.DAT. If the pertinent section of the Run
file contains the lines:
SET ^Prod1 = "PRODUCT"
SET ^Prod2 = "WT"
SET ^Date = "10/1/99"
CONNECT WEIGHT.DAT
and the SQL statement in the definition of the Connectivity Data Set contains the
statement:
SELECT ^Prod1, ^Prod2 FROM DB WHERE DATE>=#^Date#
and you start the Run File Interpreter with the command line:
QARFI.EXE $RUN=WTTASK.RUN
then Quality Analyst will take “PRODUCT” and “WT” and “10/1/99” from the
SET commands, substitute them for “^Prod1” and “^Prod2” and “^Date” in the
SQL statement, and pass the SQL statement to the database as:
SELECT PRODUCT, WT FROM DB WHERE DATE>=#10/1/99#
Configuring Connectivity
To configure Connectivity, click Settings on the main Quality Analyst start-up
screen. In the Quality Analyst Settings dialog box, select the External Database tab.
Database Options
When connecting to an external database, Quality Analyst can either open the
database, or attach the external database tables to a temporary database created by
Quality Analyst. It may seem roundabout for Quality Analyst to create a
temporary database and attach tables, but there can be advantages depending on
the particular database and type of connection used.
Note: You can override these global Database Options settings for an individual
Data Set. See Setting External Database Access Options on page 557 for
details.
Opening Databases
To open the external database select Open Database. If you choose this option,
select Pass SQL to Database for Processing if you want to pass Structured Query
Language (SQL) queries to the external database for evaluation. Clear this check
box if you want the Jet Database Engine to evaluate SQL queries.
Despite (or perhaps because of) many SQL “standards,” each database has its own
unique SQL syntax. Jet Database Engine SQL matches Microsoft Access SQL. If
the Jet Database Engine interprets the query, Microsoft Access can be used to
create the SQL statement.
If the query is passed on to the external database for evaluation, functions unique
to its SQL can be used. You can also use the pass-through setting when the table
or field names of the external database don’t match the Jet Database Engine
naming restrictions.
Attaching Databases
To attach external tables to a temporary database, select Attach Tables to
Temporary Databases.
Table and field names in remote databases don’t always match the Microsoft Jet
Database Engine syntax. In some databases, the table name has a prefix consisting
of an “owner” name and a period. Whereas the Jet Database Engine expects
database field names in the form “table.field”, these databases return
“owner.table.field.” If you select Modify Table Names for Jet Engine, Quality
Analyst will try to modify the internal table names so they are acceptable to the Jet
Database Engine (“owner.table” is changed to “owner_table”). If an SQL query is
used, it is also modified.
Note: Clearing this check box allows Quality Analyst to mimic the behavior of
earlier releases of Quality Analyst in which date (type D) variables
contained only the date portion of a date-and-time field.
Data Retrieval
In the area labeled Data Retrieval, select Snapshot to return actual data records
from an external database to Quality Analyst. Select Dynaset to return keys to
database records instead of the actual records.
Either setting will work with any access method. Which technique is fastest
depends on factors outside the scope of Quality Analyst. Only by experimenting
with both can you determine the fastest technique. The results may, in fact, be
nearly identical.
Note: You can override these global Data Retrieval settings for an individual
Data Set. See Setting External Database Access Options on page 557 for
details.
Database Debugging
In the area labeled Database Debugging, select Debug Window if you want Quality
Analyst to display a dialog box containing details of each database command
before the command is issued. This method is tedious to use, but provides
immediate feedback on each step of database connectivity.
Select Debug Logging if you want Quality Analyst to write details of each database
command to a log file.
See Debug and Message Logging on page 597 for more information.
Note: To view the log file, open the Help menu of the Quality Analyst Editor,
select About Quality Analyst, then click the Show Debug Log button. You
can print the log or copy it to the Clipboard. Northwest Analytical
technical support will often ask that you fax or e-mail them a copy of the
log (an Email button is provided for the latter). The log file is cleared
when you exit Quality Analyst.
ODBC Settings
In the area labeled ODBC Settings, enter timing parameters for ODBC (Open
DataBase Connectivity) connections.
In the Login Timeout box, enter the number of seconds Quality Analyst should
wait for the ODBC database to respond before canceling the attempt to connect to
the database.
In the Query Timeout box, enter the number of seconds Quality Analyst should
wait for the ODBC database to respond to a query before canceling the query.
In the Connection Timeout box, enter the number of seconds Windows should
keep the ODBC connection open after Quality Analyst requests it to close. The
default setting is 1 second. If you enter 0, the ODBC connection will remain open
until you close Quality Analyst. Windows does this because it’s more “efficient”
if the connection is reopened a second time. During the time Quality Analyst has
an open ODBC connection, applications that require exclusive use of the external
database cannot run. Setting this parameter to one second reduces the potential for
conflict.
settings. If strict security is vital, leave Password (and maybe even User Name)
blank and select Ask for Access Logon.
The user name and password are used whenever the Jet Database Engine is
started. Jet uses them with all external databases. In addition, with password-
protected Microsoft Access databases, the user identification and password must
have “read” permission for the database.
If you clear the Access Logon check box, the Jet Database Engine will use
“Admin” for User Name and nothing for Password.
Microsoft Access must be used to assign user identification and passwords; this
cannot be done from Quality Analyst. Once Quality Analyst has initialized the Jet
Database Engine with the user identification and password, they stay in effect
through the entire Quality Analyst session. To change the Jet Database Engine
user identification and password, you must exit and restart Quality Analyst.
External databases accessed through ODBC may have their own security systems.
Generally ODBC drivers will query you for a user identification and password. If
Microsoft Access security is used and an external database also has security, you
may be queried for several passwords before a database is available. In this
situation, the Jet engine can be run without any security so only the external
database security must be satisfied.
You can configure several security parameters to manage how Microsoft Access
security affects Quality Analyst connectivity. (See the section on MS Jet
Database Engine on page 592.) The tradeoffs between ease of use and security
should be considered for each situation.
1 NWA Quality Analyst may connect directly to the database, or ODBC can be
used (see the “Connection” list box in the External Database Setup dialog).
2 The entire database may be opened, or individual data tables may be attached
to a temporary database created by Quality Analyst (see the Open Database
setting on the External Database tab of the Quality Analyst Settings dialog
box).
3 Data tables attached to a temporary database can be renamed to match Jet
Database Engine naming restrictions (see the Modify Table Names for Jet
Engine setting on the External Database tab of the Quality Analyst Settings
dialog box).
4 Data can be retrieved from a single database table, or a user-defined SQL
query can be executed (see Specifying a Data Table on page 547).
5 A user-defined SQL query may be evaluated by the Jet Database Engine, or
passed to the external database for evaluation (see the Pass SQL to Database
for processing setting on the External Database tab of the Quality Analyst
Settings dialog box).
6 Results can be returned as keys to the data, or as the actual data records (see
the Dynaset and Snapshot settings on the External Database tab of the Quality
Analyst Settings dialog box).
When Debug Logging is selected on the External Database tab of the Quality
Analyst Settings dialog box (see page 591), all debug information is saved to a log
file. This is true even if Debug Window is not selected. In the single-user version of
Quality Analyst, the log file is named LOG.TMP and is located in the Quality
Analyst folder. In the LAN version, the log file is named LOGnnnnn.TMP (where
nnnnn is the user number) and is located in the Temporary File Path (see File
Paths and Permissions on page 269).
The debug log file is an ASCII text file and may be read by any editor or word
processor. To display the debug log file in Quality Analyst, first open the Help
menu and select About Quality Analyst.
Note: You can print the information in this dialog box or copy it to the
Clipboard by clicking Print or Copy respectively. Or you can click the
Email button to send this information directly to NWA Technical
Support.
If Debug Logging is selected, this dialog box will include a Show Debug Log
button. Click it to display the debug log file.
To print the debug information, click Print. To copy the information to the
Clipboard, click Copy. If you need technical assistance from NWA, faxing or
e-mailing this information to us may help us resolve your problem. An Email
button is provided for the latter.
APPENDIX A
Missing-Data Techniques in
Control Charts
When creating control charts, Quality Analyst recognizes and automatically
adjusts for missing data, using as much of the data as possible. Because they deal
with data in subgroups, the control-chart functions and the process-capability
functions can treat missing data in several ways.
When Quality Analyst encounters missing data, it uses one of five “missing-data
modes” to determine how to handle missing data in the control-chart and process-
capability routines. These modes are summarized in the table below. The default
missing-data mode for Quality Analyst is set in the Missing & Tagged Data tab of
the Quality Analyst Configuration dialog box as explained under Missing and
Tagged Data on page 271.
The following two sections explain how Quality Analyst handles missing data in
each mode for different subgroup types, and provides samples of each.
Summary of the Effect of Missing-Data Modes on a Sample Data File with Rowwise Subgroups
The following table shows a short example Data file with missing data. This data
will be used to illustrate the effect of missing data and missing data mode on both
types of columnwise subgroup organization.
The first set of examples below explain the effect of missing data on the analysis
and charting of subgroups comprising every three rows of data in a file. The
examples after that explain the effect of missing data on subgroups created from a
“moving window” of three rows of data.
2 4.8 1
3 5.0 1 5.10
4 3.7 2
5 4.7 2
6 * 2 *
7 3.6 3
8 3.7 3
9 3.5 3 3.60
10 * 4
11 * 4
12 * 4 *
13 5.4 5
14 4.7 5
15 5.9 5 5.33
Data X-bar
Data File Subgroup
File Chart
Row Number
Contents Values
1 5.5 1
2 4.8 1
3 5.0 1 5.10
4 3.7
5 4.7
6 *
7 3.6 2
8 3.7 2
9 3.5 2 3.60
10 *
11 *
12 *
13 5.4 3
14 4.7 3
15 5.9 3 5.33
Data X-bar
Data File Subgroup
File Chart
Row Number
Contents Values
1 5.5 1
2 4.8 1
3 5.0 1 5.10
4 3.7 2
5 4.7 2
6 * 2 4.20
7 3.6 3
8 3.7 3
9 3.5 3 3.60
10 *
11 *
12 *
13 5.4 4
14 4.7 4
15 5.9 4 5.33
Data X-bar
Data Subgroup
File Chart
File Row Number
Contents Values
1 5.5 1
2 4.8 1
3 5.0 1 5.10
4 3.7 2
5 4.7 2
6 *
7 3.6 2 4.00
8 3.7 3
9 3.5 3
10 *
11 *
12 *
13 5.4 3 4.20
14 4.7
15 5.9
Summary Overview—
Columnwise Subgroups Every 3 Rows
The following table summarizes the effect of each of the five missing-data modes
on the sample Data file with columnwise subgroups comprising every three rows
of the file.
Data Data
Subgroup
Subgroup
Subgroup
Subgroup
Subgroup
Number
Number
Number
Number
Number
Values
Values
Values
Values
Values
File File
X-bar
X-bar
X-bar
X-bar
X-bar
Chart
Chart
Chart
Chart
Chart
Row Contents
1 5.5 1 1 1 1 1
2 4.8 1 1 1 1 1
3 5.0 1 5.10 1 5.10 1 5.10 1 5.10 1 5.10
4 3.7 2 2 2 2
5 4.7 2 2 2 2
6 * 2 * 2 4.20 2 4.20
7 3.6 3 3 2 3 2 4.00
8 3.7 3 3 2 3 3
9 3.5 3 3.60 3 3.60 2 3.60 3 3.60 3
10 * 4 4
11 * 4 4
12 * 4 * 4 *
13 5.4 5 5 3 4 3 4.20
14 4.7 5 5 3 4
15 5.9 5 5.33 5 5.33 3 5.33 4 5.33
The preceding examples explain the effect of missing data on the analysis and
charting of subgroups comprising every three rows of data in a file. The following
examples explain the effect of missing data on subgroups created from a “moving
window” of three rows of data (moving average).
APPENDIX B
Alternative Statistical
Calculation Methods
three sigma limits are 0 and 6.2, while the 0.001 and 0.999 probability limits
would be set at 0 and 7.5.
With this setting, control limits will be set at the 0.001 and 0.999 probability
levels for all control charts. Warning limits will be set at the 0.025 and 0.975
probability levels for all control charts.
Process Capability
The Process Capability function can use any of the four methods for estimating
process standard deviation. The standard deviation is then used to estimate process
capability. The default method is to calculate sample standard deviation using
individual measurements.
Statistical Routines
The statistical routines in Quality Analyst (ANOVA, correlation, normal
probability plot, descriptive statistics, and summary statistics on formatted reports)
use sample standard deviation. The Descriptive Statistics functions can also use
population standard deviation if requested.
APPENDIX C
APPENDIX D
Significant Digits
One of the most troublesome areas of scientific and technical computation is the
idea of “significant digits,” or how many digits of a given number are meaningful.
Most people manage to complete their education without gaining a working
understanding of significant digits. While it is beyond the scope of this manual to
offer a detailed tutorial, we will discuss the subject as it relates to Quality Analyst.
Both impressions are incorrect. The number is represented to the limits of the
language because of the potential loss of information if rigid output formats were
used. It is your responsibility to determine the number of truly significant digits in
the output, based on the nature of your data. Some Quality Analyst routines format
certain output, when enough is known about the valid numeric range. For
example, correlation coefficients are less than or equal to 1.00, and have no more
than three significant digits. As a result, these statistics are usually limited to the
format X.XXXX when printed or displayed.
Computational Considerations
Single- and double-precision are “real” or “floating-point” number representation
systems, but they cannot take on every value in the real number continuum.
Therefore, floating-point calculations in computers are subject to round-off errors.
This is not usually critical, since errors in sequential calculations tend to cancel
each other out. However, in some cases, floating-point errors can be critical.
As you might expect, division represents the most difficult problem, since a
simple “1/3,” for example, is a repeating decimal number (0.333333...etc.).
Multiplication, which consists of repeated additions, can also generate more digits
than there were in either of the original terms. Subtracting two numbers of vastly
different size can create problems too. Finally, adding a very small number to a
very large one might produce a result identical to the initial large number.
Another possible problem involves subtracting two numbers which, if precision
were greater, would generate a zero difference, but which instead generates a very
small negative number. This can occasionally cause a subsequent computation to
fail; for example, a square root of the difference will yield the error statement
ILLEGAL FUNCTION CALL if the number is less than zero. This will usually
happen when there is a major problem with the input data, such as that which
produces a singularity in regression analysis. At any rate, you should call
Northwest Analytical if you can not find the source of the error. In most cases,
Quality Analyst will trap the most common computational problems and let you
interpret the results properly, or adjust the input data (transforming or scaling data
is a common solution).
APPENDIX E
References
There are many very good articles and books on statistics. There are even more
bad ones. The following list is not intended to be complete; however, we have
actually read and used these references, and like them.
Citations
ASTM. Manual on Presentation of Data and Control Chart Analysis: 6th Edition
(ASTM Manual Series: MNL 7). Philadelphia: American Society for Testing and
Materials 1989 (ISBN 0-8031-1289-0).
This manual presents the equations and procedures to use in developing
control charts. Includes examples, with data, of all of the charts and
calculations discussed. Also contains the table of control chart factors that
virtually all other textbooks refer to and explains how the factors are
calculated. Use this book if someone challenges your method of setting
control limits.
Burr, I.W. Engineering Statistics and Quality Control. New York: McGraw-Hill,
1953.
A good general text covering engineering statistics and control charts.
Chan, Lai K., Smiley W. Cheng, and Frederick A. Spiring. “A New Measure of
Process Capability: Cpm.” Journal of Quality Technology vol. 28 (no. 3), pp. 162-
175, July 1988.
This article explains the Cpm index used in the process capability function.
DeVor, Richard E., Tsong-how Chang, and John W. Sutherland. Statistical
Quality Design and Control: Contemporary Concepts and Methods. New York:
Macmillan Publishing Company, 1992 (ISBN 0-02-329189-X).
A good book covering control charts, process capability, and design of
experiment. Takes a more modern view of SPC with capability indices,
EWMA, and Taguchi techniques explained. Some of the more advanced
techniques are applied in an unusual fashion.
Duncan, A.J. Quality Control and Industrial Statistics 5th ed. Homewood, Illinois:
Richard D. Irwin, 1986 (ISBN 0-256-03535-0).
A very complete text that covers the statistics in detail. It tends to be more
technical than the other references.
Grant, E.L. and R. S. Leavenworth. Statistical Quality Control 6th ed. New York:
McGraw-Hill, 1988 (ISBN 0-07-024114-7).
A good working manual that concentrates on Shewhart control charts. Written
for non-statisticians, it includes many case studies. This book gives the
technical detail necessary to understand why control charts work, without
being overly dense.
Huff, Darrell How to Lie with Statistics. New York: W. W. Norton & Company,
Inc. 1954.
A somewhat irreverent classic, continuously in print since 1954 and never
revised. Covers issues important to all statistical analysis. The book’s intent is
to prevent the reader from being taken in by statistical lies, but the reverse
also works.
Hunter, J. Stuart. “The Exponentially Weighted Moving Average.” Journal of
Quality Technology vol. 18 (no. 4), pp. 203-210, October 1986.
This article explains the generation and use of the Exponentially Weighted
Moving Average (EWMA) chart. The author adds some non-standard details
such as plotting the EWMA one time position ahead, and a method for
choosing lambda. The article is complete and well written.
Ishikawa, K. Guide to Quality Control. Tokyo: Asian Productivity Organization,
1986 (ISBN 92-833-1036-5).
A simple guide to most of the control charting techniques, oriented toward the
person on the factory floor.
Johnson, N.L. and F. C. Leone. Statistics and Experimental Design in Engineering
and the Physical Sciences Vol. 1. New York: John Wiley & Sons, 1964.
Covers a wide variety of statistics useful in engineering, including control
charts. It is fairly detailed, and includes engineering examples.
Juran, J.M. (ed.) Juran’s Quality Control Handbook 4th ed. New York: McGraw-
Hill, 1988 (ISBN 0-07-033176-6).
A large (1500+ pages) collection of quality-related articles, including many
real examples from specific industries.
advanced). ASQC members receive discount prices. You should not assume,
however, that all books listed in the ASQC Catalog are worthwhile.
The ASQC toll-free order line is (800) 952-6587. The mailing address is ASQC
Quality Press, Customer Service Department, P.O. Box 3066, Milwaukee, WI
53201-3066. Their web site is at www.asqc.org.
The SPC Press, Inc. toll-free number is (800) 545-8602. Their address is
5908 Toole Drive, Suite C, Knoxville, TN 37919. Their web site is at
www.spcpress.com.
APPENDIX F
Equations
SS t = ∑ ni (X i − X )
Treatments Variation Sum of Squares t
2
i =1
Degrees of Freedom t −1
Variance: Mean S. S. (MSS) SS t
MSS t =
t −1
F ratio MSS t
F=
MSS e
SS e = ∑∑ (X ij − X i )
Error Variation Sum of Squares t ni
2
i =1 j =1
Degrees of Freedom t
∑ (n
i =1
i − 1)
∑ (n
i =1
i − 1)
SS = ∑∑ (X ij − X )
Total Variation Sum of Squares t ni
2
i =1 j =1
Degrees of Freedom t
∑n
i =1
i −1
Where:
t = number of treatments
ni = number of observations for treatment i
ni
∑X
j =1
ij
Xi = = mean of treatment i
ni
t ni
∑∑ X
i =1 j =1
ij
X = t
= grand mean
∑n
i =1
i
SS t = b∑ ( X i − X )
Treatments Variation Sum of Squares t
2
i =1
Degrees of Freedom t −1
Variance: Mean S. S. (MSS) SS t
MSS t =
t −1
F ratio MSS t
Ft =
MSS e
SS b = t ∑ ( X i − X )
Blocks Variation Sum of Squares b
2
i =1
Degrees of Freedom b −1
Variance: Mean S. S. (MSS) SS b
MSS b =
b −1
F ratio MSS b
Fb =
MSS e
SS e = ∑∑ (X ij − X i − X j + X )
Error Variation Sum of Squares t b
2
i =1 j =1
i =1 j =1
Degrees of Freedom tb − 1
Where:
t = number of treatments
b = number of blocks
b
∑X
j =1
ij
Xi = = mean of treatment i
b
∑X ij
Xj = i =1
= mean of block j
t
t b
∑∑ X
i =1 j =1
ij
X = = grand mean
tb
Chi-square Statistics
Bartlett’s Chi-square Statistic
M
χ2 =
C
Degrees of Freedom = (n − 1)
⎡ n 2 ⎤
⎢ ∑ f i si ⎥ n
( )
n
M =∑ f i ⋅ ln ⎢ i =1n ⎥ − ∑ f i ⋅ ln s i2
i =1 ⎢
⎢⎣ ∑ f i ⎥⎥ i =1
i =1 ⎦
⎡ ⎤
1 ⎢ n
1 1 ⎥
C = 1+ ⎢∑ − n ⎥
3(n − 1) ⎢ i =1 f i
⎢⎣ ∑ f i ⎥⎥
i =1 ⎦
χ2 = ∑
k
( X i − niα )2
i =1 niα (1 − α )
∑X i
α= i =1
k
∑n
i =1
i
χ2 = ∑
k
(X i − X)
2
i =1 X
χ2 = ∑
k
( X i − Ei )2
i =1 Ei
Mean X X ± A1σ
or
X ± A2 R
or
X ± Aσ ′
Standard Deviation σ′ B3σ and B4σ
or or
c2σ ′ B1σ ′ and B2σ ′
Range R D3 R and D4 R
or or
d 2σ ′ D1σ ′ and D2σ ′
Individual measurements X σ
X ±3
c2
or
R
X ±3
d2
or
X ± 3σ ′
~
Median X X ± A1σ
or
~
X ± A2 R
or
~
X ± Aσ ′
Fraction defective p p (1 − p )
p±3
n
Number defective np np ± 3 np (1 − p )
Defects per unit u u
u ±3
n
Number of defects c c ±3 c
Correlation
∑ (X − X )(Yi − Y )
n
i
rxy = i =1
∑ (X − X) ∑ (Y −Y )
n n
2 2
i i
i =1 i =1
where:
Descriptive Statistics
Arithmetic Mean (first moment n
about origin) ∑X i
X = i =1
n
Standard Deviation (sample
∑ (X − X)
n
2
statistics) i
s= i =1
n −1
Standard Deviation (population
∑ (X − X)
n
2
statistics) i
s= i =1
∑ (X − X)
Variance (sample statistics) n
2
i
s2 = i =1
n −1
∑ (X − X)
Variance (population statistics) n
2
i
s2 = i =1
n
Coefficient of Skewness m3
γ1 =
m12.5
Coefficient of Kurtosis m4
γ2 =
m22
Standard Error s
SX =
n
Q= i =1
n
Geometric Mean G = n X 1 ⋅ X 2 ⋅ ... ⋅ X n
∑ (X − X)
3rd moment about the Mean n
3
i
m3 = i =1
∑ (X − X)
4th moment about the Mean n
4
i
m4 = i =1
UCL = X + 3σˆ Zˆ
t
LCL = X − 3σˆ Zˆ
t
Kruskal-Wallis Statistic
k
R2
⋅ ∑ i − 3(n + 1)
12
H=
n(n + 1) i =1 ni
k
n = ∑ ni
i =1
N
ESS = ∑ εˆi2 = εˆ ′εˆ
i =1
εˆ = Y − Yˆ
Yˆ = Xβ
Regression Coefficients
β̂ = ( X ′X )−1 X ′Y
Standard Deviation of the Coefficients
s βi = s ( X ′X )ii−1
Estimate of variance
εˆ ′εˆ
s2 =
N −k
t-statistic for coefficients
βˆ i
t N −k =
s ( X ′X )ii−1
ANOVA Table
(
R′2 = 1 − 1 − R 2 ) NN −− k1
k = number of coefficients in model
Regression F-test
R2 N − k
Fk −1, N − k =
1− R2 k −1
Standardized (Beta) Coefficients
sX j
βˆ ∗j = βˆ j
sY
j = 2,3,..., k
Durbin-Watson Statistic
∑ (εˆ − εˆt −1 )
2
t
DW = t =2
T
∑ εˆ
t =1
t
2
USL − LSL
C pm =
6σˆ ′
where :
⎛ n ⎞
σˆ ′ = σˆ 2 + (x − m )2 ⋅ ⎜ ⎟
⎝ n −1⎠
and m = midpoint of specifications or target
When both specification limits and target are set, and the target is not at the
midpoint of the specification limits, then use:
min[T − LSL, USL − T ]
Cp =
3σ
T −μ
k=
min[T − LSL,USL − T ]
C pk = min[CPL, CPU ]
where:
T − LSL ⎛ T −μ ⎞
CPL = ⎜1 − ⎟
⎜
3σ ⎝ T − LSL ⎟⎠
USL − T ⎛ T −μ ⎞
CPU = ⎜1 − ⎟
3σ ⎜ USL − T ⎟
⎝ ⎠
Geary’s a
n xi − x
∑ n
a= i =1
n
∑ (x − x)
2
i
i =1
n
Process Capability Distribution Types
You can use several probability distributions with the process capability functions
in Quality Analyst. The probability distribution determines the shape of the
distribution curve that appears on the histogram output, and estimate of percent
outside of specifications. With all of the distribution types, Quality Analyst
assumes that the data is from a single, unimodal distribution.
Normal Distribution
Normal Distribution
1 ⎛ x−μ ⎞
2
− ⎜ ⎟
p(x ) =
1 2⎝ σ ⎠
e
σ 2π
Two Parameter Lognormal Distribution
1 ⎛ ln ( x )− μ ⎞
2
− ⎜ ⎟
p(x ) =
1 2⎝ σ ⎠
e
xσ 2π
Three Parameter Lognormal Distribution
1 ⎛ ln ( x −θ )− μ ⎞
2
1 − ⎜ ⎟
σ
p ( x) = e 2⎝ ⎠
(x − θ )σ 2π
Johnson Transformation
⎛ x −ε ⎞
S u system : p( x) = γ + η sinh −1 ⎜ ⎟
⎝ λ ⎠
⎛ x −ε ⎞
S l family : p( x) = γ + ηln⎜ ⎟
⎝ λ ⎠
⎛ x −ε ⎞
S b system : p ( x) = γ + η ⎜ ⎟
⎝ε +λ − x⎠
x = original data
γ , ε ,η, λ are estimated
Weibull Distribution
⎛ x −γ ⎞ β
β −⎜ ⎟
p(x ) = β (x − γ ) e ⎝
β −1 α ⎠
α
Rayleigh Distribution (Beta is set to 2)
2( x − γ )
⎛ x −γ ⎞
2
−⎜ ⎟
⎝ α ⎠
p ( x) = e
α2
Folded Normal Distribution
⎛ − 1 ⎛⎜ ψ ( x −γ )− μ ⎞⎟ 1 ⎛ ψ ( x −γ )+ μ ⎞ ⎞
2 2
− ⎜
p ( x) = ⎜
1
e 2⎝ σ ⎠
+ e 2⎝ σ ⎠
⎟
⎟
σ 2π ⎜ ⎟
⎝ ⎠
γ = fold point
ψ = +1 if x ≥ γ
or
ψ = −1 if x ≤ γ
−1
⎡ 1 ⎛ t −μ ⎞ ⎤
2
1 ⎛ x−μ ⎞
2
− ⎜ ⎟
⎢ 1 ⎥
− ⎜ ⎟ b 2⎝ σ ⎠
p(x ) =
1
∫e
2⎝ σ ⎠
e ⎢σ 2π dt ⎥
σ 2π ⎢ a ⎥
⎣ ⎦
a,b = lower or upper truncation point, otherwise ∞
USL − LSL
Cˆ p =
6σ̂
USL − x
Cˆ PU =
3σ̂
x − LSL
Cˆ PL =
3σ̂
USL + LSL
2 −x
2
k̂ =
USL - LSL
( )
Cˆ pk = min Cˆ PU , Cˆ PL = Cˆ p 1 − kˆ ( )
100
Cˆ R% =
Cˆ p
USL − x
Zˆ USL =
σ̂
LSL − x
Zˆ LSL =
σ̂
(
Zˆ min = min Ẑ USL ,− Zˆ LSL )
USL − LSL
Cˆ p aiag =
R
6
d2
⎛ ⎞
⎜ ⎟
⎜ USL − x x − LSL ⎟
Cˆ pk aiag = min⎜ ,
R R ⎟
⎜⎜ 3 3 ⎟
⎝ d2 d 2 ⎟⎠
USL − LSL
Pˆ p =
n
∑ (x
i =1
i − x)
2
6
n −1
⎛ ⎞
⎜ ⎟
⎜ ⎟
⎜ USL − x x − LSL ⎟
Pˆpk = min⎜ , ⎟
⎜ n n
⎟
⎜ ∑ i ( x − x ) 2
∑ ( x i − x )2 ⎟
⎜ 3 i =1 3 i =1 ⎟
⎝ n −1 n −1 ⎠
USL − LSL
Cˆ pm =
6s ′
where:
s ′ = σˆ 2 + ( x − m )
2 n
n −1
USL − LSL
m=
2
Linear y = a + bx y = a + bx
Exponential y = a ⋅ e bx ln ( y ) = ln (a ) + bx
Logarithmic y = a + b ⋅ ln (x ) y = a + b ⋅ ln (x )
Power Law y = a ⋅ xb ln ( y ) = ln (a ) + b ⋅ ln ( x )
t-tests
Paired Samples t-Statistic
n (D − μ D )
t=
sD
μ D = μ X − μY
n
∑D i
D= i =1
n
Di = X i − Yi
∑ (D − D)
n
2
i
sD = i =1
n −1
n − 3 ⎡ (1 + R )(1 − ρ 0 ) ⎤
z= ln ⎢ ⎥
2 ⎣ (1 − R )(1 + ρ 0 ) ⎦
t=
(X − Y ) − D
⎡1 1 ⎤
s 2pooled ⎢ + ⎥
⎣n m⎦
∑ (X − X ) + ∑ (Yi − Y )
n m
2 2
i
s 2pooled = i =1 i =1
n+m−2
Unequal Variance
t=
(X − Y ) − μ D
2 2
s s
+ x y
n m
μ D = μ X − μY
Welch’s Approximation of degrees of freedom for unequal Variance:
⎡ ⎤
⎢ 1 ⎥
d.f. = integer value of ⎢ ⎥
⎢ c
2
(1 − c)
2
⎥
⎢⎣ n − 1 + m − 1 ⎥⎦
s x2
c= n
2
s x2 s y
+
n m
Single-sample t-Statistic
Known Variance
n (X − μ 0 )
t=
σ
Unknown Variance
n (X − μ 0 )
t=
s
∑ (X − X )(Yi − Y )
n
s xy2 i
rxy = = i =1
∑ (X − X) ∑ (Y −Y )
sx s y n
2
n
2
i i
i =1 i =1
e 2 c2 − 1
Upper limit =
e 2 c2 + 1
INDEX
Index
$ @
$ASKFILTERn (Replaceable Parameter), @ (Run file command-line option), 503
460, 488 example, 503
$COMMENT (Run file command), 431
$DATA (Run file command-line option), 503
$DATE (title parameter), 247, 249 A
$FILE (title parameter), 247
$FILEBASE (title parameter), 247 ABS function, 25, 65
$FILEn (Replaceable Parameter), 460, 488 ACCA (Run file command), 393
$FILTER (Run file command), 439 ACCAs. See assignable causes or corrective
$MATCHFILTERn (Replaceable Parameter), actions
460, 488 accuracy, 632
$PAGEWIDTH (report parameter), 467 actions to correct violations. See corrective
$PLOT (Run file command-line option), 503 actions
$QANWA (Run file command-line option), additional limits on process capability
503 histograms, 325
$RANGEFILTERn (Replaceable Parameter), AIAG capability indices, 264, 322, 425, 445
460, 488 alphabetic variables, 44
$REPORTDATE (Run file command-line alternative calculation methods, 624
option), 503 analysis of variance. See ANOVA
$RUN (Run file command-line option), 503 analyzing
$STRINGn (Replaceable Parameter), 460, assignable causes and corrective actions,
488 148
$TEMP (Run file command-line option), 503 angle
$TIME (title parameter), 247 for rotating X-axis labels, 225
$TITLE (Run file command-line option), 503 annotating charts, 203
$VARn (Replaceable Parameter), 460, 488 ANOVA, 340
completely randomized, 340
randomized block design, 342
/ ANOVA (Run file command), 396
ANOVAR (Run file command), 397
/CLOSE (CHART subcommand), 426 APPEND (Run file command), 397
/CONTINUE (CHART subcommand), 426 Append editor, 398
/DELAY (CHART subcommand), 426 ASCII characters, 630
/PRINT (CHART subcommand), 426 AskAlpha (Run file command), 399
/SAVEAS (CHART subcommand), 426 AskAlphaMatchFilter (Run file command),
401
AskAlphaRange (Run file command), 403
AskAlphaRangeFilter (Run file command),
405
INDEX • 663
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
664 • INDEX
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
INDEX • 665
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
666 • INDEX
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
INDEX • 667
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
668 • INDEX
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
H L
labels. See description variables
Header files
LANDSCAPE (CHART subcommand), 426
creating from a Run file, 433
LastDaysFilter (Replaceable Parameter), 574
creating new, 520
LastMonthsFilter (Replaceable Parameter),
creating with Run files, 495
574
definition, 9
LikeFilter (Replaceable Parameter), 574
registration of, 15
limit regions, 98, 255, 299
updating from a Run file, 478
and Connectivity Data Sets, 307
using an existing, 519
defining, 304
histograms. See process capability histograms
synchronizing with data, 59
limits
I confidence, 178
inner, 308
importing specifying, 250
comma-separated-value files, 430, 510 warning, 308
fixed-field files, 430, 512 line thicknesses, 220
individual measurement variables, 40 for charts, 224
individuals and range charts, 129 for text reports, 224
example, 134 LN function, 25, 65
individuals charts, 129 LOG.TMP file, 597
example, 134 LOGn function, 25, 65
inner limits, 308 long file names, 392, 487
input mask Lotus Approach databases, 597
defining, 56 lower specification limit, 46
overriding, 87 LSL. See lower specification limit
inserting
data from another Data Set, 112
rows, 35, 80 M
INT function, 25, 65
managing
IR (Run file command), 442
charts, 187
Data Sets, 9
J margins
for printed charts, 195
JPEG files mask, input, 56
quality of, 228 MatchFilter (Replaceable Parameter), 574
INDEX • 669
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
670 • INDEX
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
INDEX • 671
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
672 • INDEX
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
INDEX • 673
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
674 • INDEX
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
U
T
u-charts, 141, 423
tagged data, 26 example, 144
handling, 287 units, 318
mode, 271 unequal expected frequencies Chi-square
tagging data, 81 statistic, 348
Autotag, 84 units
TAN function, 25, 65 box plots, 182, 274
target p-charts, 317
show or suppress, 318 surviving, 177
specification, 46 u-charts, 318
Task Scheduler and Run files, 498 untagging data, 81, 85
template for data entry, 56 UpdateHED (Run file command). See
temporary folders, 480 UpdateNWH
text comparisons, 249 UpdateNWH (Run file command), 478, 495
text files upper specification limit, 46
of exception report data, 163 user IDs for external databases, 589
Text Output files (obsolete), 507 USETEMPDIR (Run file command), 480,
tick marks, show minor, 282 584
tiling chart windows, 192 USL. See upper specification limit
time
database fields, 562
formats, 249 V
variables, 43
time element (in DateTime variables) variable control charts, 128
filtering, 92 examples, 131
sorting, 99 Variable Definition files (VDF), 13, 452, 538,
title 541
file, 278 creating, 539
for graphics, 246 variable description, 46
for reports, 246 variable names, 36, 62
TITLE (Run file command), 476 in calculated variable formulas, 64
title bar variable parameters, 283
variable name in, 229 variables
titles alphabetic, 44
for charts, 203 attribute, 42
ToFilter (Replaceable Parameter), 574 axis label, 57
toolbar calculated, 62
button size, 74, 192 category, 45
toggling display of, 191 data type, 37
transposing. See rotating Data Sets DateTime, 42
INDEX • 675
Northwest Analytical, Inc. NWA Quality Analyst User’s Manual
defining, 35
description, 279, 294 X
description (data type), 42
X-axis
description of, 46
labels, 281, See description variables
grouped data, 45
scale, 216
individual measurement, 40
zooming in, 214
measurement, 40
X-axis labels
multiplier, 45
rotating, 225
numeric description, 44
x-bar and range charts, 129
other, 46
example, 132
reduced data, 41
x-bar and standard deviation charts, 129
selecting for charting, 121
example, 132
selecting for report, 116
x-bar charts, 129
specifications, 46
example, 131
specifying, 284
XML files, 437
subgroup measurement, 40
Graphics files, 12
time, 43
Header file format, 113
variable-size subgroups, 23
of exception report data, 163
VDF files. See Variable Definition files
XRS (Run file command), 483
verifying data, 100
viewing
Y
chart details, 206
Graphics files, 185
V-mask, 130, 315, 435 Y-axis
scale, 216
zooming in, 214
W year 2000, 248
WALDRUN (Run file command), 481
Z
Wald-Wolfowitz runs test, 364
warning limits, 308
Weibull zone rules, 256
analysis, 174 defining, 259
and large Data Sets, 178 zones
distribution, 174 rainbow. See rainbow zones
plots, 178 zooming in on charts, 214
WEIBULL (Run file command), 482
wild cards, 9, 566
Working message
suppressing in Run file, 453
(7316)
676 • INDEX