Complete Mainframe

Mainframe Refresher Part-1
Page:1
ISPF &
TSO
Page:2
ISPF
ISPF (Interactive Spool Productivity Facility) is a popular editor in the mainframe
platform. There are two kinds of commands that can be entered in edit/view panel of
ISPF editor.
They are:
1. Primary Commands that can be entered in the Command Line.
2. Line Commands that can be entered in the line(s).
Mastering in these commands will be highly helpful in analyzing/developing
application programs.
Primary Command
KEYS
PFSHOW
RESET
X ALL
COLUMNS/COLS / COLS
ON/COL
TOP/BOT
HEX
SWAP
SWAP LIST
CUT
CUT DISPLAY
CUT APPEND
PASTE
SCRNAME XXXXX
LOCATE linenumber/Label
RECOVERY ON
RECOVERY OFF
RECOVERY OFF
UNWARN
UNDO
Meaning
Displays the PF Keys and its definition for that panel
Shows PF Keys definition at the bottom of panel. To OFF
this option enter PHSHOW OFF.
Clears all message lines and reset the display of the
dataset. RESET LABEL is required to clear labels.
Exclude all the lines
Display a ruler at the top of the screen. This can be
removed by COLS OFF or RESET.
Move to the TOP or BOTTOM of the file.
Displays the data in Hexadecimal mode. Type HEX OFF to
come back to normal display. Highly helpful for reading
computational fields and low-values.
Switch between split screens.
Lists the available screens.
With C or CC line command, this will copy the lines to
temporary storage.
CUT .A .B will cut the lines in between the labels A & B.
Display the lines already cut.
Appended the lines to the lines already cut.
Paste the already cut lines. Line commands A or B should
be used before issuing paste to indicate after or before
which line this paste should occur.
The current screen is named as XXXXX. If you are using
multiple split screens, then you can come back to this
screen by SWAP XXXXX line command. Its good
convention to name the screens with the purpose so that
you can easily come back. By default the screens are
named with numbers.
Locates the line-number or the label. Labels can be
established in any line by the line command
.labelname.
With RECOVERY on, UNDO line command will revert the
last recent change made in the edit/view session. REC
OFF will off the recovery option and UNDO Unavailable
message will appear in the first line. REC OFF UNWARN is
same as REC OFF but there wont be any UNDO
F Musa
F Musa 1
F Musa 1 30
F Musa ALL
Page:3
unavailable warning message.

Locates the next Musa.
Locates the next Musa in the first column
Locates the next Musa that appear anywhere between
column 1 and 30.
The above commands will do the find from the current
cursor position and place the cursor over the string found.
Independent of cursor position, this will count all the
Musa and displays the total count in the upper right
corner and place the cursor over the first occurrence from
top.
F Musa FIRST/LAST
Locates the First/Last Musa in the dataset/member

opened.
F Musa NEXT/PREV
Locates the Next/Previous Musa with respect to the

current cursor position in the dataset/member opened.
F Musa CHARS
Note: ALL/FIRST/NEXT/PREV/LAST are mutually exclusive

keywords.
This is same as F Musa. Look for Musa.
F Musa WORD
Look for Musa . Word is a string followed and prefixed by

space.
F Musa PREFIX/SUFFIX
Look for Musa as a prefix or suffix with some other

strings.
Note: PREFIX/SUFFIX/WORD/CHARS are mutually
exclusive.
Locates the string Musa in excluded lines.
NX searches the string only in non-excluded lines.
Note: X/NX are mutually exclusive with each other.
Locates the next occurrence of the string that was
searched in the last find command.
Changes all Musa as Muthu.
F Musa X
F * NEXT
C ALL Musa Muthu
DEL ALL X/NX

;
SORT 5 15 A
Note: Change command is nothing but find and replace. So

from string can be qualified with other keywords available
in FIND command. Ex C st1 str2 1 replaces only the str1
that are in first position to str2.
Delete all the excluded/Not excluded lines. DEL .A .B
deletes all the lines between .A and .B
Represents Enter key. So multiple commands can be given
in one shot with ; separation.
Sorts in 5-15 in ascending sequence. Maximum 5 fields can
be given. D for descending.
Page:4
Special String Characters that can be used with FIND and CHANGE:
Char
P=
Meaning
Displays any character
Char
P.
P#
P@
Displays any numeric

Displays any alphabetic
P-
P<
P>
Displays any upper-case

alphabetic
Finds the hex value of nn.
P$
Xnn
Meaning
Displays any
character
Displays any
Displays any
alphabetic
Displays any
non-displayable
non-numeric
lower-case
special character.
Displays any non-blank

character
C P 73 80 ALL changes all non-blank characters in 73-80 to spaces.

Primary commands prefixed with & (ampersand) is left in the command line after
execution. This way the same command can be entered multiple times without
retyping it in each time.
Primary Commands Mapping to PF Key
Any of the line commands can be mapped to a PF key and invoked by
pressing that key. Example if you map X ALL command with PF4, then every time
PF4 is pressed, all the lines are excluded. Mapping can be done by the line command
KEYS. In this way, 24 keys PF1-PF24 can be mapped. PF13-PF24 is nothing but Shift
+ PF1- PF12.
Typical assignment of the keys is given below:
PF1
PF2
PF3
PF4
PF5
PF6
PF7
PF8
PF9
PF10
PF11
PF12
Help
SPLIT - Split the screen at the cursors location
Exit
Exit or RETURN
RFIND - Find next occurrence of the last F command
RCHANGE Change next occurrence of the last C command
UP
DOWN
SWAP
LEFT
RIGHT
RETRIEVE - repeats the previous command line command
Page:5
LINE Commands:
Command
C
M
A/B
Rn
UC
LC
D
I
X
S
)n
(n
TS n
TE
Meaning
Copy this line. C 10 copies 10
lines. This will be followed by the
line commands A or B.
Move this line. M 10 copies
move 10 lines. This will be
followed by the line commands A
or B.
After(A)/Before(B) this line,
place the copied/moved line(s).
A10 repeats the copied/moved
lines 10 times.
Repeats the current line n times.
Just R repeat 1 time.
Change this line content to
uppercase
Change this line content to
lower-case
Deletes the line. D n deletes n
lines from this line.
Insert 1 line. After entering the
data in the first line, next line
will be inserted.
Useful for line by line data entry
ComMand
CC
Meaning
MM
Move the block of lines
RR
Repeats the block of lines. RR 10

repeats the block 10 times.
Changes the content of the
blocked lines to upper-case
Changes the content of the
blocked lines to lower-case
Deletes the block of lines.
UCC
LCC
DD
I 10
Exclude this line. X10 excludes

10 lines starting with this line.
Show one line of the excluded
text. S 10 shows 10 lines.
XX
Shift the line content by n

columns to the right side.
(outside). >n is same as )n but
in case of data truncation,
operation will be error out.
Shift the line content by n
columns to the left side.(inside)
<n is same as (n but in case of
data truncation, operation will
be error out.
To split text in two at the point
of cursor to allow a word, phrase
or sentence to be inserted. N
number of lines are inserted
between the two parts of text.
Text Entry mode is used for bulk
entry of text and allows power
typing techniques to be used at
))n
F/L
((n
TF n
BNDS
Copy the block of lines
Insert 10 lines. Once enter is

pressed, the remaining
untouched lines will be removed.
Useful for control-C, control-V
actions
Exclude a block of lines.
Shows the First/last line of
excluded text. L 10 shows the
last 10 lines.
Shift the block of lines by n
columns to the right side.
>>n is same as ))n but in case
of data truncation, operation will
be error out.
Shift the block of lines by n
columns to the left side.
<<n is same as ((n but in case
of data truncation, operation will
be error out.
To flow fragments of sentences
into paragraph. N defines the
width.
To process only over a column of
data, define the range of
columns using BNDS line
Page:6
MASK
the terminal. (Type data without

bothering about end of line, line
numbers etc .The FINAL enter
will re flow the data.)
It allows to put fixed data into
the new lines being entered.
Type MASK over a line number
and press enter. Overtype the
mask with whatever we want.
Now every time we enter new
lines with I, TE, TS mask lines
will be displayed rather than
empty lines.
command.
O
(over
lay)
Overlay command indicates the

target line(s) when merging
lines. Overlay is used in
conjunction with M or C line
commands.
C 0300 *
O 0400
MVSQuest
Result:
0300 *
0400 *
MVSQuest
O10 overlaid the first 10 lines.

OO-OO overlaid the block of
lines.
TSO
Frequently used TSO commands are listed in the table. They can be issued
from any ISPF panel with TSO prefix. TSO prefix is not required if you execute them
in TSO panel. (Option 6 of ISPF)
If the command is incomplete, TSO prompts for the missed parameters.
The output of the processed command is displayed in ISPF, in line mode.
SYS1.HELP dataset contains all the TSO commands, syntax and their usage.
Command Syntax
HELP [COMMAND]
SEND
LISTCAT
LISTDS
RENAME
Purpose
Displays the purpose and syntax of the command.
The source for this information is retrieved from SYS1.HELP.
Ex: HELP ALLOCATE
It is used to communicate with other users.
SEND REFRESHER COMPLETED USER(SMSXL86) LOGON
Message hi is sent to user SMSXL86. If the user is not logged
on, the message should be displayed the next time user logs
on. LOGON is an optional parameter.
Message: 115 chars maximum
It is used to list entries in MVS catalog. The syntax and the
available options are explained in VSAM - IDCAMS section.
LISTCAT ENTRIESSMSXL86.TEST.SOURCE ALL
It is used get the information about one or more datasets.
LISTDS dataset-name MEMBERS|HISTORY|STATUS|LEVEL
MEMBERS list all the members of a PDS. This command is
useful in REXX to process all the members of a PDS.
It is used to rename a dataset. Generic datasets are allowed.
RENAME BPAMAIN.TEST.* BPMAIN.UNIT.* renames all the
Page:7
datasets start with BPMAIN.TEST to BPMAIN.UNIT.
DELETE
ALLOCATE
(Existing dataset)
ALLOCATE
(New dataset)
FREE
ISRDDN
CALL
SUBMIT
CANCEL
STATUS
ALTER
RENAME SMSXL86.TEST(MEM1) MEM2 renames mem1 as

mem2 in the PDS SMSXL86.TEST.
It is used to delete dataset(s). PURGE qualifier is needed to
bypass expiration date checking.
DELETE BPMAIN.TEST.* PURGE
ALLOCATE DA(SMSXL86.TEST.SOURCE) FILE(INP) SHR
It allocates the dataset SMSXL86.TEST.SOURCE with logical
name INP. INP can be used in REXX for READ/WRITE.
Allocation can be done using model-dataset attributes or with
new attributes or combination of both.
Attributes that can be specified are:
UNIT, VOLUME, SPACE, TRACKS|CYLINDERS|BLOCK(blklength), DIR, RECFM, LRECL, BLKSIZE, DATACLASS,
STORCLAS, MGMTCLAS, EXPDT/RETPD.
Model parameter: LIKE(model-dataset)
De-allocate the allocated-datasets.
FREE FI(logical-name)
FREE DSNAME(dataset-name)
FREE ALL
It displays the files/datasets allocated to your session.
M member command searches the member in all the datasets
allocated to you. If you want to see the source of a utility, give
a search using the command M utility-name in the ISRDDN
panel.
It is used to execute a load module.
CALL BPMAIN.TEST.LOADLIB(TEST1) Executes the load
TEST1. If the program needs any datasets, then they should
be allocated before issuing this command.
It is used to submit a job to JES.
SUBMIT PDS(MEMBER)[ JOBCHAR(A)]
It is used to CANCEL /PURGE the job from JES.
CANCEL JOB-NAME| JOB-NAME(JOB-ID) [PURGE]
It is used to get the status of the submitted-job in JES.
STATUS JOB-NAME | JOB-NAME(JOB-ID)
It is used to alter the attributes of a dataset. Ex: Management
class.
TSO Commands can be executed in batch (JCL) using terminal monitor program
IKJEFT01.
//RUNBATCH EXEC PGM=IKJEFT01
//SYSTSPRT DD SYSOUT=*
//SYSTSIN DD *
RENAME MVSQUEST.EMP.DATA LEADSOFT.EMP.DATA
RENAME MVSQUEST.SALES.DATA LEADSOFT.EMP.DATA
/*
Page:8
The above step renames two datasets from MVSQUEST qualifier to LEADSOFT
qualifier. Any high volume manual job can be completed in matter of minutes if you
have good knowledge in ISPF and TSO commands with little exposure to REXX.
Page:9
Job Control Language

JCL
Page:10
JCL
Any business application is divided into logical modules and these modules are
developed using programming languages. These programs should be executed in a
pre-defined sequence to achieve the business functionality.
JCL (Job Control Language) is used to DEFINE and CONTROL the JOB to the
operating system.
Definition involves definition of the programs need to be executed, the data for the
programs and the sequence of programs. CONTROL involves controlling the
execution or bypassing of a program in the sequence based on the result of the prior
program execution.
JCL Coding Sheet
123
NAME
11
OPERATION
16
OPERANDS
73
80
JCL statements should have // in column 1 and 2. STAR (*) in the third column,
indicates that the line is a comment line.
NAME is optional field. If coded, it should start at column3 and can have maximum 8
characters. The first character should be an alphabet or national character (@, # or
$). Remaining characters can be any alphanumeric or national characters.
OPERATION follows NAME field. There should be at least one space between NAME
and OPERATION. If NAME is not coded then OPERATION can start at fourth column
itself. Typical OPERATION keywords are JOB, EXEC and DD.
OPERANDS are the parameters for the operation. OPERANDS follow OPERATION and
there should be at least one space between them. A comma separates parameters
and there should not any space between parameters. If the OPERANDS are more,
then they can be continued in the next line. To continue the current line, end the
current line before column 72 with , and start the next line anywhere between
columns 4-16. Columns 1-3 should be // .
COMMENT FIELD Comment field optionally follows OPERAND FIELD, preceded by at
least one blank.
End of Job is identified by NULL statement. NULL statement has // in column 1 and 2
with no NAME, OPERATION or OPERAND fields. The statements coded after NULL
statement will not be processed.
DELIMITER Some times we pass the data in the JCL itself. This is called in-stream
data. The starting of data is identified by * in the operand field of DD operation.
DELIMITER indicates the end of data. /* in column 1 and 2 is the default delimiter.
Page:11
Job Entry Subsystem (JES)

Job Entry Subsystem (JES) is the job processor of MVS operating system.
MVS installation can have either JES2 or JES3. The submitted jobs are taken by JES
for processing.
JES2
Decentralized Environment. Every
processor processes the incoming jobs
individually.
Datasets are allocated before the step
execution.
JES3
Centralized Environment. There is a
global processor that controls all the
other processors and assigns the jobs to
them.
Datasets are allocated before the job
execution.
JCL Statements
JOB. It should be the first statement in the JCL. It indicates accounting information
and JOB related information to the system. If the member being submitted contains
multiple job cards, then multiple jobs will be submitted. These jobs will run
concurrently or one after other based on job name, class and initiator availability.
EXEC. The name of the program or procedure to be executed is coded here. Every
EXEC statement in a JOB identifies one step. Maximum of 255 EXEC statements can
be coded in a JOB.
DD.
Data Descriptor. The dataset details are coded here. Dataset contains the data
that need to be processed by the program or data that is produced by the program.
Maximum 3273 DD statements can be coded in a step.
Abnormal End (ABEND) & ERROR
Once the work to be done is defined in JCL, it can be submitted to the

operating system using SUBMIT command. Usually programmer is expected to issue
JEM or JSCAN to check out any possible JCL Errors before submission.
JCL ERROR:
1. Errors before job starts execution: If there are syntax errors, then the whole
job is rejected with error message in JES MESSAGES. Typically this needs
correction and resubmission of the whole JOB.
2. Errors before step starts execution: If there is any allocation issues in a
particular (like dataset not found, duplicate dataset), then also the job will be
error out but in this case there might be already n steps got executed.
Typically this needs correction and restart in the JOB.
ABEND:
Unlike JCL Errors, ABEND happens during the execution of a program in a
step. ABENDS are classified into 2 categories.
System ABEND(Snnn): System abend occurs when the system is not able to execute
a statement that is instructed in the program. Divide by ZERO results SOCB system
abend. The OS throws it.
Page:12
User ABEND(Unnnn): When some unexpected condition occurs in the data passed,
the program will call an abend routine and abend the step with proper displays. This
is thrown by application based on the requirement.
JOB Statement
Sample Syntax:
//JOBNAME JOB (ACCOUNTING INFO), (PROGRAMMER NAME),
// TIME=(MINUTES,SECONDS), CLASS=A,MSGCLASS=A,PRTY=14,ADDR=VIRT,
// REGION=nK,MSGLEVEL=(A,B),COND=(N,OPERATOR), TYPRUN=SCAN
JOBNAME
It identifies name of the job. The job is identified in the JES SPOOL using this
name. Naming rules are already mentioned in the coding sheet section.
ACCOUNTING INFO (Mandatory. Installation Dependant.)
1. Resource usage charges are charged to the account mentioned here.
2. If you dont know your installation account, you cannot submit the job. It is like
when you dont have account, you cannot withdraw cash in bank.
3. Maximum 142 characters can be coded as accounting information.
PROGRAMMER NAME
Programmer name or program functionality or group can be mentioned. It is

used for documentation (Max 20 chars)
CLASS (Installation Defined)
1. CLASS is coded with single alphanumeric character (A-Z, 0-9).

2. During installation, every CLASS and its properties are defined.
3. Definition describes the job characteristics like CPU time usage, number of
tape/cart usage and other resource constraints.
4. Every class is assigned to one or more initiators. The jobs run in initiator address
space. One initiator can process one job at one time.
PRTY
Syntax: PRTY=N (N can be 0 15).

1. While selecting the jobs with same class for execution, JES schedules the high
priority jobs first. The job coded with PRTY=15 has the highest priority and
PRTY=0 has the lowest priority.
2. PRTY works within the JOBCLASS. If there are 2 jobs with CLASS A is submitted
and one with PRTY 3 and other with PRTY 4 then PRTY 4 will get into execution
queue first.
3. PRTY function is disabled in most of the installations.
MSGLEVEL
Syntax: MSGLEVEL=(X,Y) (X can be 0-2 & Y can be 0-1)

1. It is used to control the lists of information appear in the Job log. To get
maximum information in the listing, code MSGLEVEL as MSGLEVEL(1,1)
2. The first parameter controls the statements. (0-Only job statement, 1-JCL, JES
statements with expanded procedures, 2-Only JCL and JES statement).
3. The second parameter controls the messages. (0- Only Step execution messages,
1-All JCL, JES, operator and allocation messages).
Page:13
MSGCLASS (Installation Defined)
Syntax: MSGCLASS=X (X can be A-Z, 0-9).

1. MSGCLASS is coded with single alphanumeric character (A-Z, 0-9).
2. Each MSGCLASS is mapped to a device or location, where the messages are
routed.
ADDRSPC
It is used to specify whether the job will run in the Real storage or Virtual
storage.
Syntax: ADDRSPC={REAL|VIRT}
REAL Allocation is done in REAL storage and the program is not page-able.
VIRT Allocation is done in VIRTUAL storage and the program is page-able.
REGION
Syntax: REGION={xK | yM} (x can be 1-2096128 & y can be 1-2047).

1. It is used to specify the amount of central /virtual storage the job requires.
It can be requested in the units of kilobytes (xK) or megabytes (yM). If requested
in terms of kilobytes, then x should be multiple of 4 or the system will round it to
nearest 4K allocates for your job.
2. REGION can be coded in EXEC statement also. REGION parameter coded on JOB
card overrides the parameter coded on EXEC card.
3. Maximum virtual memory available is 2GB.
4. Region=0M allocate all the available memory in the address space to this job.
5. Region related ABENDS: When the requested region is not available, the JOB will
ABEND with S822. When the requested region is not enough for the program to
run, you will get ABEND S80A or S804.
RESTART
RESTART parameter allows restarting from any particular step in the job.
Syntax: RESTART = Step-name in the job
RESTART = * means restart from the beginning.
To restart from any procedure steps, code RESTART=PROCSTEP.STEPNAME
Whereas PROCSTEP=name of the JCL step that invoked the PROC &
STEPNAME=name of the proc step where you want execution to start.
RESTART ignores any condition in the step being restarted and it can also be step
that is in the ELSE part of the IF..ELSE..ENDIF.
TYPRUN
It is used to request special job processing.

1. TYPRUN=SCAN checks the syntax errors without actual execution.
2. TYPRUN=HOLD checks the syntax error and if there is any error, it is notified and
if there are no errors, the job is kept in awaiting execution queue and it should be
released by user for execution. Release can be done by typing A against the job
name in SDSF.
3. TYPRUN=JCLHOLD Function is same as HOLD but the syntax check starts only
after the release of the job.
Page:14
TIME
It defines the maximum allowable CPU time for the JOB. The parameter can be coded
at EXEC card also. On EXEC, it defines CPU limit of step.
Syntax: TIME = (MINUTES, SECONDS), MINUTES <= 1440 and SECONDS < 60
TIME=NOLIMIT/1440/MAXIMUM means the job can use CPU for unlimited time
TIME=0 will produce unpredictable results.
If TIME is coded on both JOB as well as EXEC, then EXEC Time limit or the
time left out in the job Time limit whichever is smaller will be the time permitted
for the step to complete.
If a JOB runs more than allowed time, then it will ABEND with system ABEND
code S322. If there is no TIME parameter, then the CPU time limit pre-defined with
CLASS Parameter will be effective.
NOTIFY
TSO User-id to whom the job END / ABEND / ERROR status should be
notified. NOTIFY=&SYSUID will send the notification to the user who submitted the
job.
COND
1. It is used for conditional execution of JOB based on return code of JOB steps.
2. The return code of every step is checked against the condition coded on JOB
card. If the condition is found TRUE, then all the steps following it are bypassed.
3. Maximum eight conditions can be coded in the COND parameter. In case of
multiple conditions, if ANY of the condition is found TRUE then the JOB stops
proceeding further.
Syntax:
COND=(CODE,OPERATOR,STEPNAME)
STEPNAME is optional. If you code it, then that particular step-name return
code is checked against the CODE with the OPERATOR. If omitted, then the return
codes of all the steps are checked. On comparison, if the condition found to be true,
then all the following steps are bypassed.
CODE can be 0-4095
OPERATOR can be GT, LT, GE, LE, EQ
It can be coded on EXEC statement. STEP level control is popular then JOB
level control. On EXEC statement, you may find ONLY, EVEN keywords against COND
parameter.
COND=ONLY allows the step execution only if any prior step is ABENDED.
COND=EVEN allows the step execution independent of any prior ABENDS.
Consider the COND parameter coded on EXEC statement,
Ex: //STEP2 EXEC PGM=PGM3,COND=((16,GE),(90,LE,STEP1),ONLY)
Step gets executed only if
A preceding step abnormally terminated OR
The return codes from all preceding steps are 17 or greater OR
The return code from STEP1 is 89 or less.
EXEC Statement
It defines the Step and Step level information to the system.

Syntax: //STEPNAME EXEC {PGM=program-name |
PROC=proc-name |
proc-name}
Page:15
STEPNAME
It is an OPTIONAL field but it is needed if you want to restart a job from this
step and for the same reason, it should be unique within the job.
PGM or PROC
Code the Program name or PROC name to be executed. (1-8 characters)

PARM
1. It is used to pass variable information to the processing program, executed by this

job step.
2. If there is more than one sub parameter or if there is any special character then
enclose them in parentheses/quotes.
3. Maximum 100 characters can be passed in this way. Quotes and brackets are not
included in this 100.To pass a quote to the program, indicate that with two quotes
one followed by other.
4. The program can receive them using linkage section. Linkage section must be
coded with half word binary field as first field. This field is populated with length of
the PARM passed from the JCL.
Example:
//STEP3 EXEC PGM=WORK,PARM=(DECK,LIST,'LINECNT=80',
//
'12+80',NOMAP)
5. To continue the PARM in the second line, start the second line from 16th position.
DPRTY
PRTY assigns priority to a job and DPRTY assigns dispatching priority to job
step. Syntax: DPRTY=(value1, value2). Value1 and value2 can be 0-15. D-Priority is
calculated using the formula (value1*16 + value)
IF /THEN/ELSE/END-IF
It is used for conditionally executing one or more steps. Nesting is possible up

to 15 levels. The meaning is same as programming IF. If the coded condition is true,
the following steps till ELSE will be executed. If the condition is false, then the steps
coded on ELSE part will be executed.
Syntax:
//name IF (relational operation) THEN
//Steps..
//
ELSE
//Steps..
//
ENDIF.
PROC PEND INCLUDE /* // //* are executed irrespective of their place.
Dont specify JOBLIB, JCLLIB, JOBCAT, STEPCAT, JOB, SYSCHK within the
THEN or ELSE Scope of IF statement.
Relational condition can be also coded as follows:
STEPNAME.ABEND=TRUE, STEPNAME.RUN=TRUE, STEPNAME NOT RUN
STEPNAME.ABENDCC = any-abend-code or
DD Statement
It defines the data requirements of the program. It is coded for every file
used in the program. If the employee details are stored in a file and catalogued with
Page:16
the name SMSXL86.EMPLOYEE.DETAILS, one of the programs (EMPPGM) in the

application reads this file, then JCL card for the DD looks like.
//STEP EXEC PGM=EMPPGM,REGION=1M
//EMPFILE DD DSN=SMSXL86.EMPLOYEE.DETAILS,DISP=SHR
When the program opens EMPFILE, it would open SMSXL86.EMPLOYEE.DETAILS.
The mapping is done as follows in the program. So in future, if you process the same
program with another file SMSXL86.EMPLOYEE.DETAILS2, then change of dataset
name in the JCL is enough. Because of logical mapping, program need not be
changed.
SELECT file1 ASSIGN TO EMPFILE. (COBOL Program)
DECLARE (EMPFILE) FILE;
(PL/I Program)
DCB DDNAME=EMPFILE
(ASSEMBLER)
FOPEN(EMPFILE,mode)
(C )
DSNAME(DSN)
The name of the dataset is coded in the DSN parameter. Dataset name can
contain 44 characters including the periods in between qualifier. Each qualifier can
have 8 characters and there can be 22 qualifiers. But usually we dont code more
than 4 qualifiers.
DSN=XXXX.YYYY.ZZZZ,DISP=SHR
Temporary datasets are indicated by && in the DSN (DSN=&&temp).
If DSN is not specified, then the system assigns the specific name to dataset.
If DSN=NULLFILE or DUMMY is coded, then all the I/O s against this file are
bypassed.
DISP
It is used to describe the status of a dataset to the system and instructs the
system what to do with that dataset after successful/unsuccessful termination of the
step or job.
DISP=(current-status, normal-termination-status, abnormal-termination-status)
Current-status can be:
NEW
SHR
OLD
MOD
Dataset does not exist. It will be created in this step

Dataset already exist and this step need it without any exclusive access
Dataset already exist and this step need it with exclusive access
If the dataset does not exist, it is to be created. If it already exists, records
are to be added to the end of the dataset. In both the cases, exclusive is
needed.
Normal-termination-status:
CATLG
DELETE
PASS
System is to place an entry, pointing to the dataset in the system/user

catalog
Dataset is no longer required. Space available for use by another dataset
but existing dataset not physically erased until overwritten by another
dataset.
Dataset is passed to subsequent steps in the same job and each step can
use the dataset only once.
KEEP
UNCATLG
Page:17
Dataset is to be kept on the volume

System is to delete the catalog entry corresponding to this dataset and
keep this dataset.
Abnormal-termination-status:
PASS is not allowed. Meaning of CATLG, UNCATLG, KEEP, DELETE are same as
normal-termination status.
Absence of any parameter should be mentioned with , as they are positional
parameters. Ex: DISP=(,,CATLG). If the dataset is a new dataset and DISP is not
coded, then the default in effect would be DISP=(NEW,DELETE,DELETE).
DCB (Data Control Block)
DCB specifies attributes of the records in the dataset.
Syntax
DCB=(LRECL=NN,BLKSIZE=YY,RECFM=Z,DSORG=MM,BUFNO=nn)
RECFM
It specifies format of the dataset. It can be Fixed, Variable, Undefined, Fixed

Blocked, variable Blocked. (F, V, U, FB, VB). Other special record formats are VBS,
FBS, VT, FT, FBA.
LRECL
It specifies logical record length. The length of the record is as in the program
for fixed length records, length of the longest record with four more bytes for
variable length records.
BLKSIZE
It contains physical record length. That is the length of the record in the
storage medium. One block contains one or more logical records. It is suggested to
code BLKSIZE as 0 so that the best size is chosen by the system, based on device.
If you explicitly code it, then it should multiple of LRECL for FB datasets and
should not be less than length of the longest records with eight more bytes for VB
dataset. In the extra eight bytes, four bytes are used for length of the record and
four bytes are used for length of the block.
DSORG.
PS (Physical sequential) PO (Partitioned organization)
BUFNO.
The number of buffers to be allocated for the dataset is coded with BUFNO
parameter. Maximum of 255 buffers can be coded. The performance of sequential
processing will increase if more buffers are allocated. The default buffers are enough
for most of the cases.
Source of DCB:
We dont always have to write the DCB parameter for a dataset. Writing DCB
parameter is one of the three ways, the information can be supplied. The other two
ways are:
1. Coded in the program. In COBOL, RECORD CONTAINS clause specifies the LRECL,
BLOCK CONTAINS clause specifies the BLKSIZE, RECORDING MODE clause specifies
Page:18
RECFM and RESERVE clause specifies BUFNO. DSORG can be assumed from the
name of the dataset and the directory space allocation of SPACE parameter.
2. Usually for an existing dataset, we dont have to code DCB parameters. It will be
available in the dataset label. The dataset label is STORED in the VTOC (DASD) or
along with dataset (TAPE) during the dataset creation.
LABEL
Syntax: LABEL = (Dataset-sequence-number
,label-type
,PASSWORD | NOPWREAD
,IN | OUT
,RETPD=nnn |EXPDT = (yyddd|yyyy/ddd))
Dataset Sequence number - identifies the relative position of a dataset on a
tape/cart volume. Should be 1 through 4 decimal digits. Omit this parameter if
access is being made to the first dataset on the tape volume.
Label - indicates the label type of the tape/cart volume.
SL - indicates that a dataset has IBM standard labels. Default value.
NL - indicates that a tape dataset has no labels.
NSL - indicates that a tape dataset has nonstandard labels.
SUL - indicates that a tape dataset has both IBM standard and user labels.
BLP - requests that the system bypass label processing for a tape dataset.
PASSWORD - indicates that a dataset cannot be read, changed, deleted or written to
unless the correct password is specified.
NOPWREAD - indicates that a dataset cannot be changed, deleted or written, unless
the correct password is specified. No password is necessary for reading the dataset.
IN - indicates that a dataset opened for I/O is to be read only.
OUT - indicates that a dataset opened for I/O is to be written only.
RETPD / EXPDT - indicates the retention period and the expiration date for a
dataset.
Ex: LABEL=EXPDT=04121 (Dataset expires on 121st day of 2004)
LABEL=RETPD=200 (Dataset is retained for 200 days)
Page:19
SPACE
It is used to request space for the new dataset. It is mandatory for all the
NEW datasets.
SPACE = ({TRK | CYL | blklgth} (,Primary-qty , Second-qty, Directory)
[,RLSE] [,CONTIG] [,MXIG] [,ROUND])
TRK/CYL - Requests that space be allocated in tracks or cylinders.
Blklgth - Specifies the average block length, in bytes, of data. Specify a decimal
number from 1 through 65535. This takes precedence, when specified, together with
the BLKSIZE field of DCB parameter.
Primary-qty - Specifies the amount of primary space required in terms of the space
unit (tracks/cylinders/number of data blocks). One volume must have enough space
for the primary quantity. If a particular volume is requested and it does not have
enough space available for the request, the job step is terminated.
Second-qty - Specifies the number of additional tracks, cylinders, blocks to be
allocated, if additional space is required.
Directory - Specifies the number of 256-byte records needed in the directory of a
PDS. (In every block we can store 5-6 members)
RLSE - requests that space allocated to an output dataset, but not used, is to be
released when the dataset is closed. Release occurs only if dataset is open for output
and the last operation was a write.
CONTIG - requests that space allocated to the dataset must be contiguous. It affects
only primary space allocation.
MIXIG It is used to specify that space requested should be allocated to the largest
contiguous area of space available on the volume. It affects only primary allocation.
ROUND - When the first parameter specifies the average block length, this parameter
requests that allocated space must be equal to an integral number of cylinders. Else
ignored.
Page:20
Extents
Extent is contiguous memory location. Only 16 extents are possible for a

physical sequential dataset in a volume. In loose terms, only 16 pointers can be
stored for a PS dataset in one volume. For a VSAM dataset it can be 123. In addition
to this, the primary (first) or secondary (consecutive) space request has to be met
within 5 extents.
If any of the above is not met, then there will be space ABEND. So for a PS
dataset, even though you request 1600 tracks (using the space parameter
(SPACE=TRKS,(100,100)), the system may not allocate you 1600 tracks always.
If all the contiguous available spaces are of size 20 tracks, then 5 extents are
used for satisfying every primary or secondary. So 15 extents are used for providing
just 300 tracks. If the system could find any 100 tracks for the 16th extent, it would
offer it and if not, there will be space ABEND. So in the best case you will get 1600
tracks and in the worst case you will get 400 tracks for the space parameter
mentioned.
Space ABENDS:
The most frequent ABEND in any production system is space abend.
SB37: End of volume. If the program tries to write more than the allocated space or
if the system could not find the requested primary or secondary space even by
joining 5 extents, then the step abnormally ends with SB37.
To solve this abend, increase the primary/secondary reasonably and if the job
again comes down, create the dataset as multi-volume by coding VOL=(,,,3). As
every volume will offer 16 extents, 48 extents in this case should be more than
enough.
SD37: If the primary space is filled and the program tries to write more but no
secondary is mentioned in the SPACE parameter, then the step will come down with
SD37.To solve this abend, provide secondary allocation in the SPACE parameter.
SE37: End of Volume. This is same as SB37. You will get this ABEND usually for a
partitioned dataset. To solve this, compress the PDS by typing Z in ISPF 3.4 panel
against the dataset or use IEBGENER. If again the job comes down, rename the old
one, reallocate the new dataset with more space and copy the old members to here
and delete the renamed dataset and restart the job.
Page:21
UNIT
It is used to request the system to place the dataset on a specific device/a

certain type or group of devices or the same device as another dataset.
UNIT=((device-number | device-type | group-name)
(,unit-count | P)
(,DEFER))
OR
UNIT=AFF=ddname
device-number - identifies a particular device by a 3-character hexadecimal number.
It should not be used unless absolutely necessary.
device-type
- Requests a device by its IBM supplied generic name. (Eg. 3380)
group-name - Requests a group of devices by its symbolic name. Installation must

have assigned the name to the device(s) during system initialization.
The group-name is 1 through 8 alphanumeric characters (Eg. TEMPDA)
unit-count - Specifies the number of devices for the dataset. (1-59)
P
- asks the system to mount all volumes for a dataset in parallel.
DEFER
- Asks the system to assign the dataset to the device but requests that
the volume(s) not be mounted until the dataset is opened. DEFER is ignored for a
new dataset on direct access.
AFF=ddname - Requests that system to allocate different datasets residing on
different removable volumes to the same device during step execution. The ddname
is that of an earlier DD statement in the same step. It reduces number of devices
used in a job step.
Ex: UNIT=(TAPE,,DEFER) UNIT=AFF=DD1
VOLUME
A reel of TAPE or a disk pack is called as one volume. VOLUME parameter is

used to identify the volume(s) on which a dataset resides or will reside.
VOLUME = ((PRIVATE)
(,RETAIN)
(,volume-sequence-number)
(,volume-count))
(SER=serial-number1,
serial number2......)
PRIVATE - Requests a private volume, that is exclusive use of volume for the dataset
specified. Only one job can access it at a time. TAPES are PRIVATE by default.
RETAIN - Requests that volume is not to be demounted or rewound after the dataset
is closed or at the end of this step. It is used when a following step is to use the
same volume.
volume-sequence-number - identifies the volume of an existing multi volume dataset
to be used to begin processing the dataset. (1-255)
Page:22
volume-count - Specifies the maximum number of volumes that an output dataset

requires. (1-255)
SER=serial-number - Identifies by serial number the volume(s) on which the dataset
resides or will reside. 1 through 6 alphanumeric or national(@,#,$) characters.
You can code a maximum of 255 volume serials.
Ex : VOLUME=SER=DEV001 VOL=(,,,3,SER=(PAGE01,PAGE02,PAGE03))
SYSOUT=class | *
It is used to identify this dataset as a system output dataset. The SYSOUT

dataset is assigned to an output class. The attributes for each class are defined
during JES initialization, including device or devices for the output class. * referbacks to MSGCLASS character of JOB CARD.
Positional and Keyword Parameters
All the parameters of JOB, DD and EXEC statements can be broadly classified
into two types. They are POSITIONAL and KEYWORD parameters.
Parameter that has its meaning defined by its position is positional parameter.
Bypassing of any positional parameter has to be informed to system by ,. Ex:
accounting information and programmer name of Job card.
Keyword parameters follow positional parameter and they can be coded in
any order. Ex: All the parameters suffixed by = are keyword parameters. PGM= and
PROC= are exceptions for this rule. They are positional parameters.
In-stream data
The data passed in the JCL stream along with JCL statements is called instream data.
Syntax: //SYSIN DD &&&&
&&&&
*
DATA
DATA,
DLM=@@
Meaning
The data follows from the next line and ends when any // or /* appears
at column 1& 2. So // and /* cannot be passed to the program.
//EMPFILE DD *
2052MUTHU
1099DEV
/*
The data follows from the next line and ends when any /* appears at
column 1 & 2. So /* cannot be passed to the program.
//SYSUT1 DD DATA
//STEP1 EXEC PGM=INV1040
//INVLSTA DD SYSOUT=A
//INVLSTB DD SYSOUT=A
/*
The data follows from the next line and ends when the characters coded
in DLM appears at column 1 & 2.
//SYSIN DD DATA,DLM=##
//EMPFILE DD *
2052MUTHU
1099DEV
/*
Page:23
##
OTHER Statements
OUTLIM
It limits the number of print lines. The limit ranges from 1 to 16777215. The
job is terminated
if the limit is reached.
//name DD SYSOUT=*,OUTLIM=3000
If the program tries to write 3001st line, JOB will ABEND with S722.
DEST
The DEST parameter is used in conjunction with the SYSOUT parameter

where the output is to be sent. This might be used where a job is run on several MVS
systems and the output is directed to a single system for the convenience of the enduser.
Syntax: //name DD SYSOUT=*,DEST=destination-ID
OUTPUT
OUTPUT statement is used to specify SYSOUT parameters for several DD

statements with a single JCL statement. It allows printing the output from single DD
statement several times, each with different SYSOUT parameters.
COPIES, OUTLIM, CLASS, DEST, FORMS, GROUPID can be coded in OUTPUT.
DEFAULT=Y can be coded on JOB and STEP level. STEP level default overrides
JOBLEVEL default.
//TEST#10 JOB ...
//STEP1 EXEC PGM=ONE
//FORM2 OUTPUT DEFAULT=YES,COPIES=2,DEST=PPP
//SYSPRINT DD
SYSOUT=A --> Produces 2 copies at PPP
//STEP2 EXEC PGM=TWO
//FORM3 OUTPUT COPIES=3,DEST=XYZ
//SYSPRINT DD
SYSOUT=Q,OUTPUT=(STEP2.FORM3,STEP1.FORM2)
--> Produces 3 copies at XYZ and 2 copies at PPP.
Maximum number of COPIES possible is 254.
FORMS
Specify the type of forms on which the SYSOUT datasets should be printed.
It is 1-8 alphanumeric or national character. SYSOUT DD FORMS parameter overrides
OUTPUT PARMS parameter.
//name OUTPUT FORMS=form-name
FREE
The datasets are allocated just before the execution of step and de-allocated
after the execution of step. FREE parameter de-allocates the file as soon as the file is
closed. //ddname DD SYSOUT=X,FREE=CLOSE
INCLUDE
The purpose of INCLUDE statement is same as COPY statement of COBOL

program. This is used to specify a PDS member that will be copied into the JCL at job
submission time. It is used to specify a standard list of DDNAMES, which would
otherwise be duplicated in many similar PROCS. This also has the advantage that
Page:24
amendments need only be made in one place. But it makes JCL unnecessarily
fragmented or difficult to read/maintain in a live environment.
//
INCLUDE MEMBER1
MEMBER1 should exist in the procedure library. Procedure libraries are coded using
JCLLIB statement. Include must not be used to execute a PROC. It is possible to nest
up to 15 levels of INCLUDE statements.
Concatenation Rules
Concatenation allows naming of more than one dataset in a single input file
without physically combining them:
//STEPLIB DD DSN=PROD.LIBRARY,DISP=SHR
//
DD DSN=TEST.LIBRARY,DISP=SHR
//
DD DSN=USR.LIBRARY,DISP=SHR
In this case Prod, Test & User libraries are concatenated.
1.16 PDS or 255 Sequential datasets can be concatenated.
2.LRECL and Record format should be same.
3.If the Block size is different, then largest block size dataset should be first.
4.Datasets may reside on different devices and device types.
REFERBACK
Page:25
The backward reference or refer back permits you to obtain information from
a previous JCL statement in the job stream. STAR (*) is the refer-back operator.
It improves consistency and makes the coding easier.
DCB, DSN, VOL=SER, OUTPUT, PGM can be referred-back.
Refer back can be done using the following ways:
1.Another DD of the same step will be referred.
*.DDNAME
2.DD of another step can be referred
*.STEPNAME.DDNAME (DDNAME of the STEPNAME)
3.DD of another proc step can be referred.
*.STEP-INVOKING-PROC.PROC-STEP-NAME.DDNAME
STAR in the SYSOUT parameter refers back to MSGCLASS of JOB card.
Refer-back example:
//STEP1 EXEC PGM=TRANS
//TRANFILE DD DSNAME=AR.TRANS.FILE,DISP=(NEW,KEEP),
//
UNIT=SYSDA,VOL=SER=MPS800,
//
SPACE=(CYL,(5,1)),
//
DCB=(DSORG=PS,RECFM=FB,LRECL=80)
//TRANERR DD DSNAME=AR.TRANS.ERR,DISP=(NEW,KEEP),
//
UNIT=SYSDA,VOL=SER=MPS801,
//
SPACE=(CYL,(2,1)),
//
DCB=*.TRANFILE
//STEP2 EXEC PGM=TRANSEP
//TRANIN DD DSNAME=*.STEP1.TRANFILE,DISP=SHR
//TRANOUT DD DSNAME=AR.TRANS.A.FILE,DISP=(NEW,KEEP),
//
UNIT=SYSDA,VOL=REF=*.STEP1.TRANFILE,
//
SPACE=(CYL,(5,1)),
//
DCB=*.STEP1.TRANFILE
.
//STEP5 EXEC PGM=*.STEP3.LOADMOD
Page:26
Special DD names
STEPLIB
It follows EXEC statement. Load modules will be checked first in this library
and then in the system libraries. If it is not found in both places, then the JOB would
ABEND with S806 code.
JOBLIB
It follows the job statement. Load modules of any steps (EXEC) that dont
have respective STEPLIB will be looked into this PDS. If not found, it will be checked
against system libraries. If it is not found there also, then the JOB would ABEND with
S806.
JCLLIB
It follows JOB statement. Catalogued procedures in the JOB are searched in

this PDS. If they are not found, they will be checked in system procedure libraries.
If they are not there, then there will be JCLERROR with Proc not found message.
Syntax: //PROCLIB JCLLIB ORDER(PDS1,PDS2)
INCLUDE members are also kept in procedure libraries. (JCLLIB)
ABEND DATASETS
In case of ABEND, one of the following three datasets will be useful. If more
than one of the three datasets is coded, then the last coded DD will be effective.
SYSUDUMP
Prints the program area, contents of registers, and gives a trace back of
subroutines called. It will be in hexadecimal format.
SYSABEND
Same as SYSUDUMP, but also prints the system nucleus. Don't use unless you
need the nucleus. It will be in hexadecimal format.
SYSMDUMP
Same information as SYSABEND, but dump will be in machine language.
Used to store dumps in a data set to be processed by an application program.
JOBCAT and STEPCAT
The datasets used in step are first checked in the STEPCAT (ICF or VSAM
Catalog) before checking in system catalog. If no STEPCAT in the step and there is a
JOBCAT, then the datasets are first searched in JOBCAT before checking in system
catalog.
SYSIN
In-stream data can be coded in SYSIN DD *. Using ACCEPT statement, these

records are read into the program. Every accept will read one line into working
storage (80 column).
Page:27
Procedures
Set of Job control statements that are frequently used are defined separately
as a procedure and it can be invoked as many times as we need from the job.
The use of procedures helps in minimizing duplication of code and probability of
error.
If a procedure is defined in the same job stream, then it is called In-stream
procedure. They are coded before the first EXEC statement in the job. The definition
starts with PROC statement and ends with PEND. Instead procedures can be saved in
a PDS and invoked from job and they are called as catalogued procedures. One
procedure can call other. This is called nesting and nesting is possible up to 15 levels.
In-stream Procedure
//JOB1 JOB
//PROC1 EXEC PROC
//STEP1 EXEC PGM=IKJEFT01
//SYSTSIN DD
DISP=SHR,DSN=MT0012.MSG(TEST)
// PEND
//STEP01 EXEC PROC1
Catalogued Procedure
//JOB1 JOB
// JCLLIB ORDER=(MT0012.PROC.PDS)
//STEP01 EXEC PROC1
MT0012.PROC.PDS(PROC1)
//STEP1 EXEC PGM=IKJEFT01
//SYSTSIN DD
DISP=SHR,DSN=MT0012.MSG(TEST)
/*
Procedure Modification
Procedure should be generic so that it can easily be used by the multiple

JOBS by simple overrides. During the invoking of procedures in the JOB, one can do
the following.
1. Override: Change the dataset names or parameters that are already coded in
the procedure
2. Addition: Add new datasets or parameters in the already existing steps of the
procedure.
3. Nullify: Omit the datasets or parameters that are already coded in procedure.
When you override a cataloged procedure, the override applies just to that execution
of the job. The cataloged procedure itself isn't changed.
Procedure Modification- EXEC statements
COND, TIME and PARM values of an EXEC statement in the procedure can be
added/modified/nullified in the invoking JCL in the following way.
//STEP1 EXEC PROC-NAME, PARAMETER-NAME.STEPNAME-IN-PROC=NEW-VALUE
Ex: PROC COBCLG has a statement
//COB EXEC PGM=IGYCRCTL,REGION=400K
// EXEC COBCLG, REGION.COB=1M
// EXEC COBCLG, TIME.COB=(0,10)
// EXEC COBCLG, REGION.COB=
=> Overrides the value of 400K.

=> Adds 10 seconds time limit for COB step.
=> Nullifies 400K region request. Default
region will be allocated now.
Page:28
Other Rules:
1. Multiple overrides are allowed but they should follow the order. That is first you
should override the parameters of step1, then step2 and then step3. Any overrides in
the wrong order are IGNORED.
2. If the STEPNAME is not coded during override, then the system applies the
override to first step alone.
//EXEC COBCLG,REGION=512K
Procedure Modification- DD Statements
DD statement in the procedure can be modified by

//STEPNAME-IN-PROC.DDNAME-OF-STEP DD parameters of dataset...
1. DD statement overrides should be done in the same order they appear in
procedure. Within a DD statement, the order of parameters does not matter.
2.Any additions should follow modifications. In a step, if you want to override the
dataset attribute of one existing dataset and add another dataset, you should
override the old one before adding the new one.
3. To omit a parameter from a DD statement in a procedure, just name it but dont
pass any value for it.
Procedure Modification Using Symbolic Parameter
A symbolic is a PROC placeholder. The value for the symbolic is supplied when
the PROC is invoked. (&symbol=value). If the value is not provided during invoke,
then the default value coded in the PROC definition would be used for substitution.
Ex: If you want to override UNIT Parameter value of all the DD statements, define
this as symbolic parameter in proc.
Catalog Procedure: PROC1
//PROC1 PROC,UNIT=SYSDA
//S1
EXEC PGM=TEST1
//DD1 DD UNIT=&UNIT
//DD2 DD UNIT=&UNIT
=> SETS Default value of UNIT as SYSDA.
//STEP1 EXEC PROC1,UNIT=TEMPDA will set &UNIT as TEMPDA for this run of
procedure.
Statements Not Allowed in a Procedure
You can place most statements in a procedure, but there are a few
exceptions. Some of these exceptions are:
1. The JOB statement and JES2/JES3 Control statements.
2. The JOBCAT and JOBLIB statement.
3. An instream procedure (an internal PROC/PEND pair)
4. SYSIN DD *, DATA statements
Nested Procedures-Add/Override/Nullification is applicable at only one level. In other
words, if PROCA calls PROCB and PROCB calls PROCC, then no statement in PROCC
can be overridden from PROCA invocation. Only PROCB can do that.
Page:29
Procedure Example
SMSXL86.TEST.PROCLIB(EMPPROC)
//EMPPROC PROC CLASS='*',SPACE='1,1'
Default values defined for CLASS
//STEP1A EXEC PGM=EMPPGM
and SPACE symbolic parameters.
//SYSOUT
DD SYSOUT=&CLASS
//EMPMAST DD DSN=&HLQ..EMPLOYEE.EDS,DISP=SHR
//
DD DSN=&HLQ..EMPLOYEE.IMR,DISP=SHR
//
DD DSN=&HLQ..EMPLOYEE.VZ,DISP=SHR
//EMPOUT
DD DSN=&&INVSEL,DISP=(NEW,PASS), INVSEL is temporary
//
UNIT=SYSDA,SPACE=(CYL,(&SPACE))
dataset
//EMPCNTL DD DUMMY
//* EMPCNTL is a control card and any in-stream data can be coded during the
//* invoke.
//*
//INV3020 EXEC PGM=EMPRPT
//SYSOUT DD SYSOUT=&CLASS
//INVMAST DD DSNAME=&&INVSEL,DISP=(OLD,DELETE)
//INVSLST DD SYSOUT=&CLASS
SMSXL86.TEST.JCLLIB(EMPJCL)
//EMPJCLA JOB (1000,200),CLASS=A,MSGCLASS=Q,NOTIFY=&SYSUID
//PROCLIB JCLLIB ORDER=(SMSXL86.TEST.PROCLIB)
// SET SPACE=1,1
Value is given for symbolic parameter SPACE.
//*STEP1A PARM is added and value for symbolic parameter HLQ is supplied.
//STEP01 EXEC EMPPROC,PARM.STEP1A=02/11/1979,HLQ=PROD
//STEP1A.EMPMAST DD
//
DD DSN=PROD.EMPLOYEE.CTS,DISP=SHR
//*Instead of PROD.EMPLOYEE.IMR, PROD.EMPLOYEE.TCS dataset is used whereas
//*other two datasets PROD.EMPLOYEE.EDS and PROD.EMPLOYEE.VZ retains their
//*position in concatenation.
//STEP1A.EMPOUT DD UNIT=TEMPDA
//*UNIT parameter of EMPOUT file is modified
//STEP1A.EMPCNTL DD *
DESIG=SSE
/*
//*EMPCNTL control card value is passed.
//STEP1A.EMPOUT2 DD DSN=PROD.EMPLOYEE.CONCAT,
//
DISP=(NEW,CATLG,DELETE),UNIT=SYSDA,
//
SPACE=(CYL,(10,10))
//*EMPOUT2 file is added to the step STEP1A.
In the above example, CLASS retains the default value coded on the PROC definition
Statement (CLASS='*').
Page:30
IEBCOPY
It is used to copy one or more members from an existing dataset to a new or

existing PDS data set. It can be also used for compressing PDS, Loading PDS to TAPE
and unloading from TAPE to disk. This utility needs two work files SYSUT3 and
SYSUT4 in addition to SYSIN and SYSPRINT.
FIELD
COPY
SELECT
EXCLUDE
LIST=YES
INDD
OUTDD
Meaning
Function is COPY
Specifies the members to be copied/replaced
Syntax: (NAME-IN-OUTPUT,NAME-IN-OUTPUT,REPLACE-IF-EXISTS)
Specifies the members to be excluded from copy
Displays the copied members in the SYSPRINT.
Points to input dataset
Points to output dataset. Should exist on the same line of COPY.
IEBCOPY- CONTROL CARD FOR MERGING TWO LIBRARIES
//SYSIN DD *
COPY OUTDD=OUTPUT INDD=(INPUT01,(INPUT02,R),LIST=NO)
/*
It says DD statements INPUT01 and INPUT02 are input files. OUTPUT is the
output file. Note the 'R' in (INPUT02,R). It instructs to IEBCOPY that like named
members are to be replaced. LIST=NO indicates that the names of the members
copied need not be listed in the SYSPRINT dataset.
IEBCOPY-CONTROL CARD FOR SELECTIVE COPY/REPLACE
COPY OUTDD=OUTPUT,INDD=INPUT01
SELECT MEMBER=((MEM1,NEWNAME,R),(MEM2,,R))
MEM1 is copied as NEWMEM in OUTPUT. If already NEWMEM exist, it will be replaced.
IEBCOPY-CONTROL CARD FOR OMITTING SELECTED MEMBERS
COPY OUTDD=OUTPUT,INDD=INPUT01
EXCLUDE MEMBER=(MEM1,MEM2)
All the members except MEM1 and MEM2 are copied into OUTPUT from INPUT01.
IEBCOPY-Complete step for Compressing PDS
//COMPRESS EXEC PGM=IEBCOPY

//SYSPRINT DD SYSOUT=*
//COMPFILE DD DSN=MM01.COPYLIB.COB,DISP=OLD
//SYSUT3 DD UNIT=SYSDA,SPACE=(TRK,(1,1))
//SYSUT4 DD UNIT=SYSDA,SPACE=(TRK,(1,1))
//SYSIN
DD *
COPY OUTDD=COMPFILE,INDD=COMPFILE
/*
Page:31
IEBGENER
In addition to SYSIN and SYSPRINT datasets, it needs SYSUT1 and SYSUT2

datasets. SYSUT1 is coded with input dataset and SYSUT2 is coded with output
dataset. If attributes were not given for SYSUT2, then the program would assume
SYSUT1 attributes for SYSUT2.
It is primarily used as COPY utility. If you want to copy any TAPE file to DISK
or DISK to TAPE, then no SYSIN is needed.
If you want to reformat your input file or if you want to create members out
of your PS file, then you need control card (SYSIN) and the first statement should be
GENERATE.
FIELD
GENERATE
MAXNAME
MAXGPS
MAXFLD
MAXLITS
MEMBER
RECORD
IDENT
RECORD
FIELD
Meaning
First Statement which sets the values for MAXNAME,MAXGPS,
MAXLITS, MAXFLDS
Maximum MEMBER statements that can follow.(During member
generation)
Syntax: MAXNAME=3
Maximum IDENT statement that can follow. (During member
generation)
Maximum FILED statements that can follow. (During reformatting)
Syntax: MAXFLDS=10
Maximum size of literal during reformatting.
It identifies the name of the member to be created.
Syntax: MEMBER NAME=MEM1
It usually follows MEMBER statement to identify the last record to be
copied from the input dataset.
RECORD IDENT= (Length,Literal,Start-Column)
Example: RECORD IDENT=(3,MVS,1), then the last record to be
copied into the member from the input dataset, has MVS in column
1-3.
It is used for reformatting the records in the input file.
RECORD FIELD=(Length, literal or input column, conversion, output
column)
Output column says where the field should be placed in the output
file. Conversion can be ZP or PZ. PZ means the input packed decimal
field is being converted into zoned format and ZP is the reverse.
IEBGENER- SYSIN CARD FOR CREATING THREE MEMBERS FROM INPUT PS FILE
//SYSIN DD *
GENERATE MAXNAME=3,MAXGPS=2
MEMBER NAME= MEMB1
RECORD IDENT=(8,'11111111'.1)
MEMBER NAME=MEMB2
RECORD IDENT=(8,'22222222',1)
MEMBER NAME=MEMB3
//
IEBGENER creates three members. It reads input file writes into memb1 until it finds
11111111 in column 1. In the same way it reads and writes the records into memb2
Page:32
until it finds 22222222 in column 1. The remaining records in the input dataset are
copied into MEMB3.
IEBGENER- SYSIN CARD FOR REFORMATTING DURING COPY
//SYSIN
DD *
GENERATE MAXFLDS=5,MAXLITS=4
RECORD FIELD=(5,1,,1),FIELD=(20,21,,6),FIELD=(9,61,ZP,26),
FIELD=(9,70,ZP,31),FIELD=(4,'TEST',,36)
/*
Input Column
Values in column 1-5
Values in 70-9
Any Conversion
Convert the zoned into

packed before copying.
Convert the zoned into
packed before copying.
Output column
Copied into column 1-5
Copied into column 6-25
Packed value is written in
26-30
Packed value is written in
31-35
TEST literal is written in
column 36-39
IEHLIST
It is used to list
1. The entries in the catalog. (SYSIN KEYWORD- LISTCTLG)
2. Directory(s) of 1-10 PDS (SYSIN KEYWORD- LISTPDS)
3. The entries in VTOC.
(SYSIN KEYWORD-LISTVTOC)
Code SYSIN, SYSPRINT and one more DD that will mount the volume queried in
SYSIN.
The following JOB lists the VTOC of INTB01 in a formatted way.
//MYJOB JOB CLASS=A,MSGCLASS=A,REGION=256K,MSGLEVEL=(1,1)
//LISTVTOC EXEC PGM=IEHLIST
//VOLDD
DD UNIT=SYSDA,VOL=SER=INTB01,DISP=OLD
//SYSIN
DD *
LISTVTOC FORMAT,VOL=3330=INTB01
/*
To list the contents of any PDS:
LISTPDS DSNAME=(SYS1.LINKLIB), VOL=SER=INTB01.
To list the catalog for a specific DASD volume:
LISTCTLG VOL=3350=PUB000
IEHMOVE
It is used to move one dataset from one volume to another volume.

//STEP01 EXEC PGM=IEHMOVE
//FROMVOL DD VOL=SER=TST001,DISP=OLD,UNIT=SYSDA /*ALLOCAT FROM VOLUME*/
//TOVOL
DD VOL=SER=PRD001,DISP=OLD,UNIT=SYSDA /*ALLOCATE TO VOLUME*/
//DD01
DD UNIT=SYSDA,VOL=REF=SYS1.SVCLIB,DISP=OLD
//SYSIN
DD *
MOVE PDS=MUTHU.TEST.PDS,TO=3380=PRD001,FROM=3380=TST001
Page:33
/*
FROM clause in the SYSIN is not needed for catalogued datasets. It is suggested to
allocate the Supervisor Call Library.
IEBCOMPR
It is used to compare two PS or PDS datasets. Two PS are same, if the

number of records is same and all the records are identical. SYSIN is not needed for
PS comparison. If they are not identical, then the following will be listed in the
SYSPRINT.
DD statements that define the dataset, Record and Block numbers, the
unequal records and maximum of 10 unequal records found.
Two PDS are same, if the corresponding members contain same number of
records and all the records are identical. SYSIN should have COMPARE TYPE=PO for
PDS.
//SYSUT1 INPUT DATASET 1
//SYSUT2 INPUT DATASET 2
//SYSPRINT
//SYSIN DD *
IEBBTPCH
IEBEDIT:
One typical interview question is how to run the selected steps. For example, how to
execute step4 and step9 of 10 steps JCL. The typical answer is to restart the job
from step4 and include a ALWAYS TRUE condition (like COND=(0,LE) or
COND=(4096,GT)) in steps 5,6,7,8 and 10. If the interviewer said COND should not
used, then only way is IEBEDIT.
//M665235C JOB (MVSQuest),'IEBEDIT TEST',
//
CLASS=B,MSGCLASS=X,NOTIFY=V665235,REGION=28M
//*
//SUBMIT EXEC PGM=IEBEDIT
//SYSUT1 DD DSN=TEST.MUTHU.JCL(JCLINP),DISP=SHR
//SYSUT2 DD SYSOUT=(*,INTRDR)
//SYSIN DD *
EDIT START=M665235C,TYPE=INCLUDE,STEPNAME=(STEP0004,STEP0009)
//*
In the above JCL, JCLINP is the 10 steps JCL. M665235C is the job-name in the JCL.
If TYPE is exclude, then the mentioned steps will not be copied/submitted.
DFSORT
If you do a global search of your JCL inventory, you will find the program that
is used very frequently is SORT. There are two famous SORT products are available in
the market. One is DFSORT and the other is SYNCSORT. The basic commands in both
the products are same.
Page:34
ICETOOL provides a lot more than what SORT can offer and it comes with
DFSORT product. SYNCTOOL comes with SYNCSORT product. PGM=SORT can point
to DFSORT or SYNCSORT. It is actually an alias to SORT product in your installation.
DFSORT is IBM product and it needs the following datasets for its operation.
SORTIN (Input dataset), SORTOUT (Output dataset), SYSIN (Control Card) and
SYSOUT (Message dataset).
Message dataset can be altered using MSGDDN= parameter of SYSIN.
SORT Card to copy all the records from SORTIN to SORTOUT
SORT FIELDS=COPY.
SORT card to skip first 100 records and then copy 20 records
SORT FIELDS=COPY SKIPREC=100 STOPAFT=20

SORT Card to sort the records based on key fields
SORT FIELDS=(STARTPOS,LENGTH,TYPE,ASC|DESC)
Type = CH (Character), BI (Binary), ZD (Zoned Decimal), PD(Packed Decimal),
FS (Signed numbers)
Ex: SORT FIELDS=(1,10,CH,A,15,2,CH,A)
SORTS all the SORTIN records with 1-10th column as major key and 15-16th column
as minor key before writing to SORTOUT.
SORT card to select the records meeting the criteria
INCLUDE COND=(STARTPOS,LENGTH,TYPE,RO,VALUE)
RO-Relational operator can be EQ,NE,LT,GT,LE,GE.
Card to select the records with TRICHY in the column 4-9
INCLUDE COND= (4,6,CH,EQ,CTRICHY)
Card to select the records which has same values in 2-3 and 5-6
INCLUDE COND= (2,2,CH,EQ,5,2,CH)
SORT card to reject the records meeting the criteria
OMIT COND=(STARTPOS,LENGTH,TYPE,RO,VALUE)
Card to reject the records with TRICHY in the column 4-9
OMIT COND= (4,6,CH,EQ,CTRICHY)
Card to reject the records which has same values in 2-3 and 5-6
OMIT COND= (2,2,CH,EQ,5,2,CH)
SORT card to change PD to ZD
If input file has a PD field S9(5)V99 Comp-3 and to reformat as PIC S9(5).9(2) then
use,
OUTREC FIELDS=(1,5,PD,EDIT(STTTTT.TT),SIGNS=(,-,,))
SORT card to remove the duplicates
SORT FIELDS= (1,5,CH,A),EQUALS
SUM FIELDS=NONE.
SORTIN records are sorted on the key 1-5 and if more than one record is
found to have same key, then only one record will be written to SORTOUT. If EQUALS
is coded, then the record to be written is the FIRST record else it can be anything.
SORT card to sum the values for same-key records
SORT FIELDS= (1,5,CH,A),EQUALS

SUM FIELDS=(10,5,PD)
Page:35
SORTIN records are sorted on key 1-5 and if more than one record is found to
have same key, then the records are summed on column 10-14 and one record is
written with total sum.
SORT card to add sequence number to the output file
OUTREC=(1,20,SEQNUM,4,ZD) 4 digit zoned decimal sequence number is
added with all the records of input file in column 21-24
This will be useful for reading a file from bottom to top. This will be useful for
matching logic in JCL. Matching logic in JCL will be explained later.
SORT card to restructure the input file before feeding to sort
INREC FIELDS=(37,2,6,6,40,4,31,2)
The length of the output file is 14.
SORT card to create multiple files from single input file (Maximum 32 files)
OUTFIL FILES=1 INCLUDE=(1,6,CH,EQ,CMUMBAI)
OUTFIL FILES=2 INCLUDE=(1,6,CH,EQ,CTRICHY)
Code output files as SORTOF1 and SORTOF2.
SORT card to restructure the sorted
OUTREC FIELDS=(1:1,20,
21:CMUTHU,
26:10Z,
36:21,10)
file before writing

=> FIRST 20 CHAR FROM INPUT FILE
= > FOLLOWED BY STRING MUTHU
=> FOLLOWED BY 10 BINARY ZEROS
=> 21ST to 10 CHARACTERS FROM INPUT FILE.
SORT card to change any XX in the column 6-7 to YY

OUTREC FIELDS=(1:1,5,
6:1,2,CHANGE=(2,C'XX',C'YY'),NOMATCH=(6,2),
8,42)
SORT card to merge
MERGE FIELDS=(STARTPOS,LENGTH,TYPE,ASC|DESC,STARTPOS,)
128 such Keys can be given. Datasets to be merged are coded in SORTIN00 to
SORTIN99.
SORT CARD to extract all the PROCS executed in a JCL
OPTION COPY
INCLUDE FORMAT=SS,COND=(1,81,EQ,C'EXEC',AND,1,81,NE,C'PGM=)
ICETOOL
DD statements in ICETOOL:
TOOLMSG
FOR
ICETOOL MESSAGES
DFSMSG
FOR
SORT MESSAGES
TOOLIN
FOR
ICETOOL-CONTROL-CARD
XXXXCNTL
FOR
SORT-CONTROL-CARD USED BY ICETOOL
XXXX is coded in USING clause of TOOLIN.
TOOLIN card to copy
COPY FROM(INDD) TO(OUTDD)
( Up-to 10 DD can be given).
Page:36
TOOLIN card to copy unique/selected duplicate records
SELECT FROM(INDD) TO(OUTDD) ON (STARTPOS,LENGTH,TYPE)

NODUPS/ALLDUPS/LOWER(n)/HIGHER(n)/EQUAL(n)/FIRST/LAST
NODUPS
ALLDUPS
HIGHER(n)
LOWER(n)
EQUAL(n)
FIRST
LAST
COPY only the unique records.

COPY only the duplicates.
COPY those duplicates that occurs more than n times (n => 1-99)
COPY those duplicates that occurs less than n times (n => 1-99)
COPY those duplicates that occurs exactly n times (n => 1-99)
- Retains the first copy in case of duplicates
- Retains the first copy in case of duplicates
TOOLIN Card to get the statistics of a numeric field
STATS FROM(INDD) ON(START,LENGTH,TYPE)

Print the maximum, average and total for numeric fields. (10 ON possible)
TOOLIN Card to get the number of unique values in a field
UNIQUE FROM(INDD) ON(START,LENGTH,TYPE)

Print count of unique values.
TOOLIN Card to get all the values for a particular field
DISPLAY FROM(INDD) ON(STARTPOS,LENGTH,TYPE) LIST(LISTDD)

Prints values of a field in the input dataset to LISTDD. (20 ON possible)
TOOLIN Card to get all the values for a particular field With Occurrence constraint
OCCURS FROM(INDD) ON(STARTPOS,LENGTH,TYPE) LIST(LISTDD) OPTION

OPTION = > HIGHER(n) LOWER(n) EQUAL(n) ALLDUPS NODUPS
HIGHER(2) means only the values that are repeated more than 2 times is reported at
LISTDD dataset.
TOOLIN Card to get number of records fell into the range mentioned
RANGE FROM(INDD) ON(START,LENGTH,FORMAT) LIST(OUTDD) options

Options are OCCURS HIGHER(n)/LOWER(n)/HIGHER(n1) LOWER(n2)/
EQUAL(n) NOTEQUAL(n)
It prints count of records meeting value criteria and the FORMAT should be numeric.
TOOLIN card to invoke SORT
SORT FROM(INDD) TO(OUTDD) USING(xxxx)

SORT statements should be coded under the DDNAME xxxxCNTL
Matching Logic in JCL
I have two files file1 and file2. I want to generate three reports out of these
two files.
1.The first report should have records that exist in both files.
2.The second report should contain records that exist only in first file and not in
second file.
3.The third report should contain records that exist only in the second file and not in
the first file.
//STEP0100 EXEC PGM=ICETOOL
//*
//TOOLMSG DD SYSOUT=*
Page:37
//DFSMSG DD SYSOUT=*
//IN1
DD *
1234567890
3456789012
5678901234
//IN2
DD *
3456789012
7890123456
8901234567
//T1
DD DSN=&T1,SPACE=(CYL,(5,5),RLSE),DISP=(,PASS)
//T2
DD DSN=&T2,SPACE=(CYL,(5,5),RLSE),DISP=(,PASS)
//INT
DD DSN=*.T1,DISP=(OLD,PASS),VOL=REF=*.T1
//
DD DSN=*.T2,DISP=(OLD,PASS),VOL=REF=*.T2
//FILEA
DD SYSOUT=*
//FILEB
DD SYSOUT=*
//OUT
DD SYSOUT=*
//TOOLIN DD *
SORT FROM(IN1) USING(CTL1)
SORT FROM(IN2) USING(CTL2)
SORT FROM(INT) USING(CTL3)
//CTL1CNTL DD *
SORT FIELDS=(1,10,CH,A)
OUTFIL FNAMES=T1,OUTREC=(1,80,C'1')
//CTL2CNTL DD *
OUTFIL FNAMES=T2,OUTREC=(1,80,C'2')
//CTL3CNTL DD *
SUM FIELDS=(81,1,ZD)
OUTFIL FNAMES=OUT,INCLUDE=(81,1,ZD,EQ,3),OUTREC=(1,80)
OUTFIL FNAMES=FILEA,INCLUDE=(81,1,CH,EQ,C'1'),OUTREC=(1,80)
OUTFIL FNAMES=FILEB,INCLUDE=(81,1,CH,EQ,C'2'),OUTREC=(1,80)
/*
Explanation:
CTL1 Add 1 to all the records of the first file at 80th column
CTL2 Add 2 to all the records of the second file at 80th column
CTL3 Concatenate both files and sort the file on key if duplicates found, sum on
81st column. So if any record exists in both the file, it will have 3 after summing.
So now extract records with 1 , 2 and 3 into three files. While writing the
records, remove the 81st byte added for our temporary purpose.
1 Records only in first file
2 Records only in second file.
3 Records exist in both the files.
IEHPROGM
It is used to
1.Catalog a dataset (CATLG DSNAME=A.B.C, VOL=SER=nnnn)
2.Uncatalog a dataset (UNCATLG DSNAME=A.B.C)
3.Rename a dataset (RENAME DSNAME=A.B.C,VOL=SER=nnnn,NEWNAME=D.E.F)
4.Create an index for GDG (BLDG INDEX=gdg-name, LIMIT=n, [,EMPTY][,DELETE])
Page:38
5.Deleting the index for GDG (DLTX INDEX=index-name)

The SYSIN cards are given in bracket. The utility needs two work datasets and
SYSPRINT for messages. Continuation of control card needs to be indicated by X in
72nd column.
If your shop installed SMS, then uncatalog wont work it out because SMS
handles the catalog.
IEHINITT
It is used to initialize a tape. It will write the volume serial number to the tape.
//STEP01 EXEC PGM=IEHINITT
//LABEL DD DCB=DEN=3,UNIT=(TAPE,,DEFER),DISP=(,KEEP),LABEL=(,SL)
//SYSIN DD*
LABEL INITT SER=DEVSMS,DISP=REWIND
/*
Generation Data Group (GDG)
GDG is group of datasets that are related to each other chronologically or

functionally. Each of these dataset is called a generation. The generation number
distinguishes each generation from others.
If the GDG Base is MM01.PAYROLL.MASTER, then their generations are
identified using the generic name MM01.PAYROLL.MASTER.GnnnnVxx.
nnnn is generation number (01-9999) and xx is version number (00-99).
Referring Generations in JCL
The current generation is referred by GDGBASE(0), the previous generation

by GDGBASE(-1) and the next generation by GDGBASE(+1).
GENERATIONS ARE UPDATED ONLY AT THE END OF THE JOB. It means, if the
first step creates one generation, code it as GDGBASE(+1) and if the second step
creates another generation, then it SHOULD be coded as GDGBASE(+2) as the (+1)
version is not yet promoted to current version. Similarly to refer the GDG created in
the second step, refer it by GDGBASE(+2).
GDG datasets can be also referenced with their generation number like
MM01.PAYROLL.MASTER.G001V00
Advantage of GDG
1. GDG datasets are referred in the JCL using GDG base and relative number. So the
same JCL can be used again and again without changing the dataset name and this
is the biggest advantage of GDG.
2.GDG Base has pointers to all its generations. When you want to read all the
transactions done till today, you can easily do it by reading the GDG base if it is
available. Otherwise you have to concatenate all the transaction files before reading.
Creation of GDG
1.GDG Base is created using IDCAMS. The parameters given while creating the GDG
are:
Page:39
Parameter
NAME
LIMIT
EMPTY/NOEMPTY
SCRATCH/
NOSCRATCH
OWNER
FOR DAYS (n) /
TO (DATE)
Purpose
Base of the GDG is given here.
The maximum number of GDG version that can exist at any point
of time. It is a number and should be less than 256.
When the LIMIT is exceeded,
EMPTY keeps ONLY the most recent generation.
NOEMPTY keeps the LIMIT number of newest generation.
SCRATCH un-catalogue and deletes the versions that are not kept.
NOSCRATCH just does un-cataloguing and it is not physically
deleted from the volume.
Owner of the GDG.
Expiry date. Can be coded either in the unit of days or till
particular date.
2. Model dataset is defined after or along with the creation of base. Once model DCB
is defined, then during the allocation of new versions, we no need to code DCB
parameter. Model DCB parameter can be overridden by coding new parameter while
creating the GDG version. It is worth to note that two GDG version can exist in two
different formats.
A step that defines a GDG and allocates a model DSCB
//GDG
EXEC PGM=IDCAMS
//MODEL DD DSNAME=MM01.PAYROLL.MASTER,DISP=(,KEEP),
//
UNIT=SYSDA,VOL=SER=MPS800,SPACE=(TRK,(0)),
//
DCB=(DSORG=PS,RECFM=FB,LRECL=400)
//SYSIN DD *
DEFINE GDG ( NAME(MM01.PAYROLL.MASTER) LIMIT(5)
NOEMPTY
SCRATCH )
/*
How to Solve ABENDS?
There are two kinds of abends- USER and SYSTEM.
USER (Unnnn)
Prefixed with U
Application Driven The application
program issues the user abend by calling
installation specific abend routine.
IBM Supplied abend routine: ILBOABN0
MOVE 999 TO ABEND-CODE
CALL ILBOABN0 USING ABEND-CODE
This code Will abend the program with
U0999.
SYSTEM (Snnnn)
Prefixed with S
System Driven When the system is not
able to perform a statement, it abended
with respective system abend.
Page:40
Ex: In a billing application, the country

referred in the bill-to address is missing
in the country table which is an serious
issue and so the program will be written
in such a way to abend in this scenario.
Solution: Look for the abend code in the
program and study the application logic
behind this abend. Then appropriately fix
the data/rerun.
Ex: Trying to do an arithmetic operation

on non-numeric data. As system could
not perform this, it will abend with S0C7.
Solution: Refer IBM Manuals for system
abend and take appropriate action. You
may need to analyse program/data to fix
based on type of abend.
System abends can be environment (space/region) related or program

related(logic/data driven logic). Fixing a program related abend is comparatively
complicated than enviroment related system abends or user abends.
The usual process to solve a system abend is:

1.Refer the SYSOUT of the job and get the next sequential instruction to be executed
(Offset).
2.Compile the program with LIST option if the compiled one is not with either LIST or
OFFSET option.
3.Check for the offset in the compilation list and Get the respective statement
number.
4.Identify the statement. This would be a numeric operation on non-numeric data.
5.Identify the source of the non-numeric data and correct it.
Though the process looks simple and this will work 90 percent of the cases, the
questions are:
In case offset and the abending module is not displayed in the sysout, how to
proceed?
In case of data-exception kind of abend, if the particular statement referred
more than one field, how do you conclude which field have problem without display
and rerun?
If the source for the field in problem is in file, how do you know which record
of this field in the file?
Dump Reading exposure will help you in all these cases. As dump reading knowledge
is important for any maintenance / support project, let us study with a simple
example:
Simple Program
CBL LIST
IDENTIFICATION DIVISION.
PROGRAM-ID.SANSOFT.
*
ENVIRONMENT DIVISION.
*
DATA DIVISION.
WORKING-STORAGE SECTION.
Page:41
01 WS-VARIABLES.
05 WS-EMP-NAME
05 WS-EMP-AGE
05 WS-EMP-CITY
05 WS-EMP-SAL
05 WS-EMP-BONUS
05 WS-EMP-CTC
PIC
PIC
PIC
PIC
PIC
PIC
X(10).
9(02).
X(10).
S9(08).
S9(08).
S9(08).
*
PROCEDURE DIVISION.
*
MOVE 'MUTHU'
TO WS-EMP-NAME
MOVE 29
TO WS-EMP-AGE
MOVE 'TRICHY'
TO WS-EMP-CITY
MOVE 60000
TO WS-EMP-SAL
COMPUTE WS-EMP-CTC = (WS-EMP-SAL * 12)
DISPLAY 'SANSOFT COMPLETED'
DISPLAY 'EMPLOYEE DETAIL:' WS-EMP-NAME
WS-EMP-AGE
WS-EMP-CITY
WS-EMP-SAL
WS-EMP-CTC
STOP RUN.
+ WS-EMP-BONUS
','
','
','
','
The program is compiled with LIST option to get assembler listing.

On successful compilation, the program is submitted with the JCL.
The program is abended and the sysout says:
CEE3207S The system detected a data exception (System Completion Code=0C7).
From compile unit SANSOFT at entry point SANSOFT at compile unit
offset +0000036A at entry offset +0000036A
at address 2790117A.
The instruction at the Offset 36A is failed. So look into the compilation listing for the
statement that is in the offset 36A.
Page:42
000021 MOVE
000358 D207
000022 COMPUTE
00035E F247
000364 D20F
00036A FA54
000370 940F
000374 F844
00037A F374
000023 DISPLAY
2016 A0BD
MVC
22(8,2),189(10)
(BLW=0)+22
PGMLIT AT +185
D0F8
D0E8
D0F2
D0F3
D0F3
2026
PACK
MVC
AP
NI
ZAP
UNPK
248(5,13),30(8,2)
232(16,13),141(10)
242(6,13),248(5,13)
243(13),X'0F'
243(5,13),243(5,13)
38(8,2),243(5,13)
TS2=0
TS1=0
TS1=10
TS1=11
TS1=11
WS-EMP-CTC
WS-EMP-BONUS
PGMLIT AT +137
TS2=0
201E
A08D
D0F8
D0F3
D0F3
TS1=11
TS1=11
35E-37F belongs to the COMPUTE statement that is in line 000022.

Now look for line 00022 in compilation listing.
000021
000022
000023
MOVE 60000
TO WS-EMP-SAL
COMPUTE WS-EMP-CTC = (WS-EMP-SAL * 12) + WS-EMP-BONUS
DISPLAY 'SANSOFT COMPLETED'
So one of this field referred in this statement has junk in it. Just before compute we
populated WS-EMP-SAL and so there is a problem with WS-EMP-BONUS. If you go
thru the code, you will find the developer missed to populate/initialize WS-EMPBONUS and that has caused data exception.
If these fields are from file, we cannot easily confirm like above. So we have to give
display for these two fields in the program and rerun the program or look for junks in
the source file for these two fields using FILE AID/ INSYNC. The other approach will
be look into data division map in the compilation listing.
0Source
Hierarchy and
Base
Hex-Displacement
Asmblr Data
Data Def
LineID
Data Name
Locator
Blk
Structure
Definition
Data Type
Attributes
2 PROGRAM-ID SANSOFT---------------------------------------------------------*
8
1 WS-VARIABLES. . . . . . BLW=00000 000
DS 0CL46
Group
9
2 WS-EMP-NAME . . . . . BLW=00000 000
0 000 000
DS 10C
Display
10
2 WS-EMP-AGE. . . . . . BLW=00000 00A
0 000 00A
DS 2C
Disp-Num
11
2 WS-EMP-CITY . . . . . BLW=00000 00C
0 000 00C
DS 10C
Display
12
2 WS-EMP-SAL. . . . . . BLW=00000 016
0 000 016
DS 8C
Disp-Num
13
2 WS-EMP-BONUS. . . . . BLW=00000 01E
0 000 01E
DS 8C
Disp-Num
14
2 WS-EMP-CTC. . . . . . BLW=00000 026
0 000 026
DS 8C
Disp-Num
WS-EMP-SAL is in BLW-0 offset 16 whereas WS-EMP-BONUS in offset 01E.

In the dump, look for working-storage dump. BLW-0 maps to 279890B8
address.
WORKING-STORAGE for SANSOFT
BLW-0: 279890B8
+000000 279890B8 D4E4E3C8 E4404040 4040F2F9 E3D9C9C3
C8E84040 4040F0F0 F0F6F0F0 F0C00000
|MUTHU
29TRICHY
0006000...|
+000020 279890D8 00000000 00000000 00000000 00000000
00000000 00000000 00000000 00000000
|................................|
BLW-0: 279890B8
+000000 279890B8
+000020 279890D8
D4E4E3C8 E4404040 4040F2F9 E3D9C9C3

00000000 00000000 00000000 00000000
C8E84040 4040F0F0 F0F6F0F0 F0C00000

00000000 00000000 00000000 00000000
|MUTHU
29TRICHY
0006000...|
|................................|
Page:43
WS-EMP-NAME is 10 bytes starting at 0 - D4E4E3C8E44040404040 which is

nothing but MUTHU
. WS-EMP-AGE is 2 bytes starting at 10th byte which is 29
(F2F9). Similarly read the contents of WS-EMP-CITY, WS-EMP-SAL, WE-EMP-BONUS
and WS-EMP-CTC.
WS-EMP-BONUS contains low-values. It is a computational field and so it is
expected to have sign in the last nibble whereas it doesnot as it is not properly
initialised in the program.
In the dump, in every line there will be twenty bytes hexa decimal content will
be followed by character content. Due to the column limit, I have the one line dump
in three lines.
So if there are million records in a file and during the processing it abended
after n number of records, to identify the record caused problem, read the dump for
the file section unique variable(s) values and look for the respective record in the file
and analyse/correct/delete.
We have taken offset directly from sysout. If it is not available, then refer the
PSW. Based on AMODE 24/31, the last 24/31 bits contain the next sequential
instruction to be executed. From this value, subtract the entry point of the program
being abended and that will give you offset. One instruction above this offset is the
one that caused the abend. Entry point of all the programs executed be found in
trace back or save trace section of the dump. You can directly found the offset in the
traceback also. (last 31 bits of A7901180 is 27901180, 27901180-27900E10 = 370
is the offset, from the compilation listing one instruction just before 370 is 36A.)
PSW..... 078D1000 A7901180
Traceback:
DSA Addr
0001D120
0001D018
Program Unit
CEEHDSP
SANSOFT
Statement
PU Addr
057780F0
27900E10
PU Offset
+00003C34
+0000036A
Load Mod
Service
Status
CEEPLPKA
SANSOFT
UK00165
Call
Exception
Entry
CEEHDSP
SANSOFT
E Addr
057780F0
27900E10
E Offset
+00003C34
+0000036A
Submission of Job from COBOL Program

Write the Job statements to a output file and route the file to INTRDR instead
of cataloguing. INTRDR-Internal Reader
JOB: //STEP1 EXEC PGM=MAINPGM
//DD1
DD DSN=MUTHU.TEST,DISP=SHR
//JCLDD DD SYSOUT=(*,INTRDR)
PROGRAM:MAINPGM
SELECT JCLFILE ASSIGN TO JCLDD. (Environment Division)
FD JCLFILE.
01 JCL-REC PIC X(80).
(File Section)
OPEN OUTPUT JCLFILE.
(Open in output and write JCL statements)
MOVE '//TESTJOB JOB 1111'
TO JCL-REC.
Page:44
MOVE '//STEP01 EXEC PGM=IEFBR14' TO JCL- REC

CLOSE JCLFILE
(TESTJOB will be submitted automatically)
Submission of Job from CICS Program
JCL can be submitted from CICS by SPOOL OPEN, SPOOL WRITE and SPOOL
CLOSE commands. This needs SPOOL=YES entry in SIT table. Refer CICS section for
more details.
Storage Management Subsystem
It is optional feature of the operating system that manages user datasets. The
SMS parameters apply only to the new datasets and the system ignores for existing
datasets, DD* or DD DATA datasets, JOBCAT and JOBLIB DD statements, and
SYSOUT datasets.
The following data classes establish various default values for catalogued
datasets. An administrator assigns a name to each group of default values, and then
you reference this name on your DD statements to use the values.
STORCLAS Defines UNIT and VOL parameters
DATACLAS - Defines RECFM, LRECL, AVGREC, SPACE, RETPD/EXTPD, DSNTYPE,
KEYLEN, RECORG, KEYOFF..etc
MGMTCLAS Migration, backup, releasing unused space in a dataset.
If you want to override any one of the values of default, you can do that.
//PDS DD DSN=BPMAIN.MUTHU.SOURCE,DISP=(NEW,CATLG),
//
STORCLAS=DASDONE,SPACE=(,(,,50)),DATACLAS=COB2
Overrides the directory space defined for COB2 data class.
JCL for SAR-IN-BATCH

The production Job logs are usually routed to one of the third party products.
SAR is one such product. The other products are CA-VIEW and VIEW-DIRECT.
The following JCL is used to query the SAR in batch.
//LOADLOG EXEC PGM=SARBCH
//STEPLIB DD DSN=SYSP.CAI.CAILIB,DISP=SHR
//REPORT
DD DSN=CORP.DBSY2KT.JOBRPT,DISP=MOD
//LOADDD
DD DSN=CORP.DBSY2KT. LOADLOG,DISP=OLD
//SYSIN
DD DSN=CORP.DBSY2KT.UTILS(CYCCARDW),DISP=SHR
To load the complete log (JOB PRODBKUP of generation 4941) into a dataset
named as LOADDD, use the following card:
/LOAD DDNAME=LOADDD ID=PRODBKUP GEN=4941
To get run-date, run-time, return code and generation of all the prior runs of
a job, use the following card. The result will be stored in the dataset named as
REPORT.
/LIST ID=JOBNAME
Page:45
JCL for CA7-IN-BATCH

CA7 is a scheduler product. We brief about the product in the next section.
This JCL is used to query CA7 in batch.
//UCC77BTI PROC ID=0,POOL='1-8',DSNPFX=,OPT=,PG=SASSBSTR
//BTERM
EXEC PGM=SAASBSTR,
//
PARM='0,POOL=(1-8) '
//STEPLIB
DD DISP=SHR,DSN=CAI.CA7.LOADLIB
//UCC7CMDS DD DISP=SHR,DSN=CAI.CA7.COMMDS
//SYSPRINT DD SYSOUT=*,DCB=BLKSIZE=133
//ERRORS
DD SYSOUT=*,DCB=BLKSIZE=133
//SYSUDUMP DD SYSOUT=*
//SYSIN DD*
Code the CA7 commands
/*
JCL FOR XDC-IN-BATCH
There may be a need to read the run details of a job log inside the REXX.
This can be done using the OUTPUT command of TSO. The following JCL captures the
complete log information of the job CENNAEFD with ID JOB08079 into
T.ISP.MUTHU(TEST1).
//RUNTSO EXEC PGM=IKJEFT01
//SYSTSIN DD *
OUTPUT CENNAEFD(JOB08079) PRINT('T.ISP.MUTHU(TEST1)')
/*
//*
Page:46
Common Business Oriented Language

COBOL
Page:47
COBOL (COmmon Business Oriented Language)

History.
Developed in 1959 by a group called COnference on DAta Systems Language
(CODASYL). First COBOL compiler was released in 1960.
First ANSI approved version
1968
Modified ANSI approved version 1974 (OS/VS COBOL)
Modified ANSI approved version 1985 (VS COBOL 2)
This book is written based on IBM COBOL for OS/390 V2R2.
Speciality.
1. First language developed for commercial application development, which can
efficiently handle millions of data.
2. Procedure Oriented Language
- Problem is segmented into several tasks.
Each task is written as a Paragraph in Procedure Division and executed in a
logical sequence as mentioned.
3. English Like language
Easy to learn, code and maintain.
Coding Sheet.
1
12
COL-A
72
80
COLUMN-B
1.6
Sequence number Should be in sequence but need not be consecutiveUsually (1-3) columns identify page number and (4-6) columns identify line
number.
7
Continuity (-), Comment (*), Starting a new page (/) and Debugging lines (D)
8-11 Column A Division, Section, Paragraph, 01, 77 declarations must begin here.
12.72 Column B All the other declarations/statements must begin here.
73-80 Identification field. It will be ignored by the compiler but visible in the source
listing.
Language Structure.
Character
Word
Clause
Statement
Sentence
Paragraph
Digits (0-9), Alphabets (A-Z), Space (b),

Special Characters (+ - * / ( ) = $ ; > < . ,)
One or more characters - User defined or Reserved
One or more words. It specifies an attribute for an entry
One or more valid words and clauses
One or more statements terminated by a period
One or more sentences.
Page:48
Section
Division
Program
One or more paragraphs.

One or more sections or paragraphs
Made up of four divisions
Divisions in COBOL.
There are four divisions in a COBOL program and the data division is an optional one.
1.Identification Division.
2.Environment Division.
3.Data Division.
4.Procedure Division.
Identification Division.
This is the first division and the program is identified here. Paragraph PROGRAM-ID
followed by user-defined name is mandatory. Though 30 characters can be entered
for the program ID, compiler will consider only the first EIGHT characters and the
remaining characters will be ignored. All other paragraphs are optional and used for
documentation.
PROGRAM-ID.
PROGRAM NAME.
AUTHOR.
COMMENT ENTRY.
INSTALLATION.
COMMENT ENTRY.
DATE-WRITTEN.
COMMENT ENTRY.
DATE-COMPILED.
COMMENT ENTRY.
SECURITY.
COMMENT ENTRY.
Security does not pertain to the operating system security, but the
information that is passed to the user of the program about the security features of
the program.
Environment Division.
This is the only machine dependant division of COBOL program. It supplies
information about the hardware or computer equipment to be used on the program.
When a program is moved from one computer to another computer, the only section
that may need to be changed is ENVIRONMENT division.
Configuration Section.
It supplies information about the computer on which the program will be compiled
(SOURCE-COMPUTER) and executed (OBJECT-COMPUTER). It consists of three
paragraphs SOURCE COMPUTER, OBJECT-COMPUTER and SPECIAL-NAMES.
This is OPTIONAL section from COBOL 85.
SOURCE-COMPUTER. IBM-4381 (Computer and model # supplied by manufacturer)
WITH DEBUGGING MODE clause specifies that the debugging lines
in the program (statements coded with D in column 7) should also be compiled and
included in load module.
OBJECT-COMPUTER. IBM-4381
(Usually same as source computer)
Page:49
SPECIAL-NAMES. This paragraph is used to relate hardware names to user-specified

mnemonic names.
1. Substitute character for currency sign. (CURRENCY SIGN IS litearal-1)
2. Comma can be used as decimal point. (DECIMAL-POINT IS COMMA)
3. Default collating sequence can be changed. It will be explained later.
4. New class can be defined using CLASS keyword. (CLASS DIGIT is 0 thru 9)
Input-Output Section.
It contains information regarding the files to be used in the program and
consists of two paragraphs FILE-CONTROL & I-O CONTROL.
FILE CONTROL. Files used in the program are identified in this paragraph.
I-O CONTROL. It specifies when check points to be taken and storage areas that are
shared by different files.
Data Division.
Data division is used to define the
It has three sections.
FILE SECTION
WORKING-STORAGE SECTION
LINKAGE SECTION
data that need to be accessed by the program.

describes the record structure of the files.
is used to for define intermediate variables.
is used to access the external data.
Ex: Data passed from other programs or from
PARM of JCL.
Literals, Constants, Identifier,

1. Literal is a constant and it can be numeric or non-numeric.
2. Numeric literal can hold 18 digits and non-numeric literal can hold 160 characters
in it. (COBOL74 supports 120 characters only)
3. Literal stored in a named memory location is called as variable or identifier.
4. Figurative Constant is a COBOL reserved word representing frequently used
constants. They are ZERO/ZEROS/ZEROES, QUOTE/QUOTES, SPACE/SPACES,
ALL, HIGH-VALUE/HIGH-VALUES, LOW-VALUE/LOW-VALUES.
Example: 01 WS-NAME PIC X(04) VALUE DEVI.
In this declaration DEVI is non-numeric literal and WS-NAME is an identifier or a
variable.
Declaration of variable
Level# $ Variable $ Picture clause $ Usage Clause $ Sync clause $ Value clause.
FILLER
Level#
It specifies the hierarchy of data within a record. It can take a value from the
set of integers between 01-49 or from one of the special level-numbers 66 77 88
01 level.
Specifies the record itself. It may be either a group item or an

Elementary item. It must begin in Area A.
02-49 levels. Specify group or elementary items within a record. Group level items
must not have picture clause.
66 level.
77 level.
88 level.
Page:50
Identify the items that contain the RENAMES clause.

Identify independent data item.
Condition names.
Variable name and Qualifier

Variable name can have 1-30 characters with at least one alphabet in it.
Hyphen is the only allowed special character but it cannot be first or last letter of the
name. Name should be unique within the record. If two variables with same name
are there, then use OF qualifier of high level grouping to refer a variable uniquely.
Ex: MOVE balance OF record-1 TO balance OF record-2.
FILLER
When the program is not intended to use selected fields in a record structure, define
them as FILLER. FILLER items cannot be initialized or used in any operation of the
procedure division.
PICTURE Clause
Describes the attributes of variable.
Numeric
Numeric Edited
Non Numeric
Exclusive sets
9 (Digit), V (Implied decimal point), S (Sign), P(Assumed

decimal point when the point lies outside the data item)
+ (Plus Sign), - (Minus Sign), CR DB (Credit Debit Sign)
. (Period), b (Blank), ,(comma), 0 (Zero), / (Slash)
BLANK WHEN ZERO (Insert blank when data value is 0),
Z (ZERO suppression), * (ASTERISK), $(Currency Sign)
A (alphabet), B (Blank insertion Character), X(Alpha numeric),
G(DBCS)
1. + - CR DB
2. V .
3. $ + - Z * (But $ Can appear as first place and * as
floating. $***.**)
DBCS (Double Byte Character Set) is used in the applications that support large
character sets. 16 bits are used for one character. Ex: Japanese language
applications.
VALUE Clause
It is used for initializing data items in the working storage section. Value of item
must not exceed picture size. It cannot be specified for the items whose size is
variable.
Syntax:
VALUE IS literal.
VALUES ARE literal-1 THRU | THROUGH literal-2
VALUES ARE literal-1, literal-2
Literal can be numeric without quotes OR non-numeric within quotes OR figurative
constant.
SIGN Clause
Syntax
SIGN IS (LEADING) SEPARATE CHARACTER (TRAILING).
It is applicable when the picture string contain S. Default is TRAILING WITH NO
SEPARATE CHARACTER. So S doesnt take any space. It is stored along with last
digit.
Page:51
+1=A +2=B +3=C +4=D +5=E +6=F +7=G +8=H +9=I

-0=}, -1= J, -2= K, -3=L, -4=M, -5=N, -6=O, -7=P, -8=Q, -9=R
Number
-125
+125
TRAILING SIGN
(Default)
12N
12E
LEADING SIGN
J25
A25
LEADING
SEPARATE.
-125
+125
Page:52
Refreshing Basics
Nibble.
04 Bits is one nibble. In packed decimal, each nibble stores one digit.
Byte.
08 Bits is one byte. By default, every character is stored in one byte.
Half word.
16 Bits or 2 bytes is one half word. (MVS)
Full word.
32 Bits or 4 bytes is one full word. (MVS)
Double word. 64 Bits or 8 bytes is one double word. (MVS)
Usage Clause
DISPLAY
Default. Number of bytes required equals to the size of the data item.
COMP
Binary representation of data item.
PIC clause can contain S and 9 only.
S9(01) S9(04)
Half word.
S9(05) S9(09)
Full word.
S9(10) - S9(18)
Double word.
Most significant bit is ON if the number is negative.
COMP-1:
Single word floating point item. PIC Clause should not be specified.
The sign is contained in the first bit of the of the leftmost byte and the
exponent is contained in the remaining 7 bits of the first byte. The last
3 bytes contain the mantissa.
COMP-2:
Double word floating-point item. PIC Clause should not be specified.
7 bytes are used for mantissa and hence used for high precision
calculation.
COMP-3:
Packed Decimal representation. One digit takes half byte.
PIC 9 (N) comp-3 data item would require (N + 1)/2 bytes. The sign is
stored separately in the rightmost half-byte regardless of whether S is
specified in the PICTURE or not.
C Signed Positive D Signed Negative F-Unsigned Positive.
INDEX
It is used for preserve the index value of an array. It takes 4 bytes.
PIC Clause should not be specified. When the clause is specified for a
group item, it applies to all elementary items contained in it. However,
the group itself is not an index data item.
POINTER
4 Byte elementary item that can be used to accomplish limited base
addressing. It can be used only in SET statement, Relation condition,
USING phrase of a CALL statement, an ENTRY statement or the
procedure division statement. A value clause for a pointer data item
can contain only NULL or NULLS.
SET identifier-1 TO address of identifier-2
SET ADDRESS OF identifier-2 TO identifier-1.
Identifier-1 is POINTER data item and identifier-2 is linkage section
item.
Page:53
SYNC Clause and Slack Bytes

SYNC clause is specified with COMP, COMP-1 and COMP-2 items. These items are
expected to start at half/full/double word boundaries for faster address resolution.
SYNC clause does this but it may introduce slack bytes (unused bytes) before the
binary item.
Consider the following declaration
01 MY-DATA.
05
DATA-ONE
05
DATA-TWO
05
DATA-THREE
in the WORKING-STORAGE SECTION:

PIC
PIC
PIC
X(6).
9(6)
S9 (4) V99
COMP SYNC.
COMP-3.
The Starting address of Full-word should end with 0,4,8 or C and that of half-word
should end with 0,2,4,6,8,A,C,E. If DATA-ONE starts at 0, it will occupy 0-5 bytes in
memory. DATA-TWO - a sync item of full word cannot start at 6. So by SYNC rule, it
starts at 8th position. 6th & 7th bytes are unused. So MY-DATA occupies 16 bytes.
REDEFINES
The REDEFINES clause allows you to use different data description entries to
describe the same computer storage area. Redefining declaration should immediately
follow the redefined item and should be done at the same level. Multiple redefinitions
are possible. Size of redefined and redefining need not be the same. It cannot be
done at 66 and 88 levels.
Example:
01 WS-DATE
PIC 9(06).
01 WS-REDEF-DATE REDEFINES WS-DATE.
05
WS-YEAR
PIC 9(02).
05
WS-MON
PIC 9(02).
05
WS-DAY
PIC 9(02).
RENAMES
It is used for regrouping of elementary data items in a record. It should be declared
at 66 level. It need not immediately follows the data item, which is being renamed.
But all RENAMES entries associated with one logical record must immediately follow
that record's last data description entry. RENAMES cannot be done for a 01, 77, 88
or another 66 entry. It cannot be done for occurrences of an array.
01 WS-REPSONSE.
05 WS-CHAR143
PIC X(03).
05 WS-CHAR4
PIC X(04).
66 ADD-REPSONSE RENAMES WS-CHAR143.
CONDITION name
It is identified with special level 88. A condition name specifies the value that a field
can contain and used as abbreviation in condition checking.
01 SEX PIC X.
88 MALE
VALUE 1
88 FEMALE VALUE 2 3.
IF SEX=1 can also be verified as IF MALE in Procedure division.
SET FEMALE TO TRUE moves value 2 to SEX. If multiple values are coded on
VALUE clause, the first value will be moved when it is set to true.
Page:54
JUSTIFIED RIGHT
This clause can be specified with alphanumeric and alphabetic items for right
justification. It cannot be used with 66 and 88 level items.
OCCURS Clause
OCCURS Clause is used to allocate physically contiguous memory locations to
store the table values and access them with subscript or index. Detail explanation is
given in Table Handling section.
LINKAGE SECTION
It is used to access the data that are external to the program. JCL can send
maximum 100 characters to a program thru PARM. Linkage section MUST be coded
with a half word binary field, prior to actual field. If length field is not coded, the first
two bytes of the field coded in the linkage section will be filled with length and so
there will be last 2 bytes data truncation in the actual field.
01 LK-DATA.
05 LK-LENGTH
PIC S9(04) COMP.
05 LK-VARIABLE
PIC X(08).
LINKAGE section of sub-programs will be explained later.
Page:55
Procedure Division.
This is the last division and business logic is coded here. It has user-defined sections
and paragraphs. Section name should be unique within the program and paragraph
name should be unique within the section.
Procedure division statements are broadly classified into following categories.
Statement Type
Imperative
Conditional
Compiler Directive
Explicit Scope
terminator
Implicit Scope
terminator
Meaning
Direct the program to take a specific action.
Ex: MOVE ADD EXIT GO TO
Decide the truth or false of relational condition and based on
it, execute different paths.
Ex: IF, EVALUATE
Directs the compiler to take specific action during compilation.
Ex: COPY SKIP EJECT
Terminate the scope of conditional and imperative statements.
Ex: END-ADD END-IF END-EVALUATE
The period at the end of any sentence, terminates the scope of
all previous statements not yet terminated.
MOVE Statement
It is used to transfer data between internal storage areas defined in either file section
or working storage section.
Syntax:
MOVE identifier1/literal1/figurative-constant TO identifier2 (identifier3)
Multiple move statements can be separated using comma, semicolons, blanks or the
keyword THEN.
Numeric move rules:
A numeric or numeric-edited item receives data in such a way that the
decimal point is aligned first and then filling of the receiving field takes place.
Unfilled positions are filled with zero. Zero suppression or insertion of editing
symbols takes places according to the rules of editing pictures.
If the receiving field width is smaller than sending field then excess digits, to
the left and/or to the right of the decimal point are truncated.
Alphanumeric Move Rules:
Alphabetic, alphanumeric or alphanumeric-edited data field receives the data
from left to right. Any unfilled field of the receiving filed is filled with spaces.
When the length of receiving field is shorter than that of sending field, then
receiving field accepts characters from left to right until it is filled. The
unaccomodated characters on the right of the sending field are truncated.
When an alphanumeric field is moved to a numeric or numeric-edited field,
the item is moved as if it were in an unsigned numeric integer mode.
CORRESPONDING can be used to transfer data between items of the same
names belonging to different group-items by specifying the names of group-items to
which they belong.
Page:56
MOVE CORRESPONDING group-1 TO group-2
Group Move rule

When MOVE statement is used to move information at group level, the
movement of data takes place as if both sending and receiving fields are specified as
alphanumeric items. This is regardless of the description of the elementary items
constituting the group item.
Samples for understanding MOVE statement (MOVE A TO B)
Picture of A
Value of A
Picture of B
PIC 99V99
PIC 99V99
PIC 99V999
PIC9(05)V9(03)
PIC 9(04)V9(02)
PIC 99V99
PIC X(04)
12.35
12.35
12.345
54321.543
23.24
00.34
MUSA
PIC 999V99
PIC 9999V9999
PIC 9V99
PIC 9(03)V9(03)
PIC ZZZ99.9
PIC $$$.99
XBXBXB
Value of B after
Move
012.35
0012.3500
2.34
321.543
23.2
$.34
MUS
ARITHMETIC VERBS
All the possible arithmetic operations in COBOL using ADD, SUBTRACT,
MULTIPLY and DIVIDE are given below:
Arithmetic Operation
ADD A TO B
ADD A B C TO D
ADD A B C GIVING D
ADD A TO B C
SUBTRACT A FROM B
SUBTRACT A B FROM
C
SUBTRACT A B FROM
C GIVING D
MULTIPLY A BY B
MULTIPLY A BY B
GIVING C
DIVIDE A INTO B
DIVIDE A INTO B
GIVING C
DIVIDE A BY B
GIVING C
DIVIDE A INTO B
GIVING C
REMAINDER D
A
A
A
A
A
A
A
B
A+B
B
B
A+B
B-A
B
C
C
A+C
A+B+C+D
A+B+C
A
A
A*B
B
A*B
A
A
B/A
B
B/A
A/B
Integer (B/A)
C-(A+B)
C-(A+B)
Integer
remainder
GIVING is used in the following cases:

1.To retain the values of operands participating in the operation.
Page:57
2.The resultant value of operation exceeds any of the operand size.
ROUNDED option
With ROUNDED option, the computer will always round the result to the
PICTURE clause specification of the receiving field. It is usually coded after the field
to be rounded. It is prefixed with REMAINDER keyword ONLY in DIVIDE operation.
ADD A B GIVING C ROUNDED.
DIVIDE..ROUNDED REMAINDER
Caution: Dont use for intermediate computation.
ON SIZE ERROR
If A=20 (PIC 9(02)) and B=90 (PIC 9(02)), ADD A TO B will result 10 in B
where the expected value in B is 110. ON SIZE ERROR clause is coded to trap such
size errors in arithmetic operation.
If this is coded with arithmetic statement, any operation that ended with SIZE
error will not be carried out but the statement follows ON SIZE ERROR will be
executed.
ADD A TO B ON SIZE ERROR DISPLAY ERROR!.
COMPUTE
COMPUTE statement assigns the value of an arithmetic operation (on the right
hand side) to a data item (on the left hand side).
Rule: Left to right 1.Parentheses
(( ))
2.Exponentiation
(**)
3.Multiplication and Division (* and /)
4.Addition and Subtraction (+ and -)
Caution: When ROUNDED is coded with COMPUTE, some compiler will do rounding
for every arithmetic operation and so the final result would not be precise.
All arithmetic operators have their own explicit scope terminators. (END-ADD,
END-SUBTRACT, END-MULTIPLY, END-DIVIDE, END-COMPUTE). It is suggested to use
them.
CORRESPONDING is available for ADD and SUBTRACT only.
INITIALIZE
VALUE clause is used to initialize the data items in the working storage
section whereas INITIALIZE is used to initialize the data items in the procedure
division.
INITIALIZE sets the alphabetic, alphanumeric and alphanumeric-edited items
to SPACES and numeric and numeric-edited items to ZERO. This can be overridden
by REPLACING option of INITIALIZE. FILLER, OCCURS DEPENDING ON items are not
affected.
Syntax:
INITIALIZE identifier-1
REPLACING (ALPHABETIC/ALPHANUMERIC/ALPHA-NUMERIC-EDITED
NUMERIC/NUMERIC-EDITED)
DATA BY (identifier-2 /Literal-2)
Example:
01
A.
Page:58
05
A1
PIC
9(5).
05
A2
PIC
X(4).
INITIALIZE A REPLACING NUMERIC DATA BY 50. will initialize only A1 by 50.
ACCEPT
ACCEPT can transfer data from input device or system information contain in
the reserved data items like DATE, TIME, DAY.
ACCEPT WS-VAR1 (FROM DATE/TIME/DAY/OTHER SYSTEM VARS).
If FROM Clause is not coded, then the data is read from terminal. At the time
of execution, batch program will ABEND if there is no in-stream data from JCL and
there is no FROM clause in the ACCEPT clause.
DATE option returns six digit current date in YYMMDD
DAY returns 5 digit current date in YYDDD
TIME returns 8 digit RUN TIME in HHMMSSTT
DAY-OF-WEEK returns single digit whose value can be 1-7 (Monday-Sunday
respectively)
DISPLAY
It is used to display data. By default display messages are routed to SYSOUT.
Syntax:
DISPLAY identifier1| literal1 (UPON mnemonic name)
STOP RUN, EXIT PROGRAM & GO BACK
STOP RUN is the last executable statement of the main program. It returns
control back to OS.
EXIT PROGRAM is the last executable statement of sub-program. It returns
control back to main program.
GOBACK can be coded in main program as well as sub-program as the last
statement. It just gives the control back from where it received the control.
ALTER statement
The alter statement is used to modify the targets of GO TO statements
written elsewhere in the procedure division.
ALTER PROCEDURE-NAME-1 TO [PROCEED TO] PROCEDURE-NAME-2
[PROCEDURE-NAME-3 TO {PROCEED TO} PROCEDURE-NAME-4 ]....
Each of the PROCEDURE-NAME-1, PROCEDURE-NAME-3 is the name of the
paragraph that contains only one sentence. This sentence must consist of a single
GO TO statement without the depending clause.
During the execution each of the PROCEDURE-NAME-1, PROCEDURE-NAME-3,
. . . will be replaced by PROCEDURE-NAME-2, PROCEDURE-NAME-4 ...respectively.
Collating Sequence
There are two famous Collating Sequence available in computers. IBM and
IBM Compatible machine use EBCDIC collating sequence whereas most micro and
many mainframe systems use ASCII collating sequence. The result of arithmetic and
alphabetic comparison would be same in both collating sequences whereas the same
is not true for alphanumeric comparison.
EBCDIC (Ascending Order)
ASCII (Ascending Order)
Page:59
Special Characters
a-z
A-Z
0-9
Special Characters
0-9
A-Z
a-z
Default collating sequence can be overridden by an entry in OBJECT-COMPUTER and

SPECIAL NAMES paragraphs.
1. Code the PROGRAM COLLATING SEQUENCE Clause in the Object computer
paragraph. PROGRAM COLLATING SEQUENCE IS alphabet-name
2. Map the alphabet-name in the SPECIAL-NAMES paragraph as follows:
ALPHABET alphabet-name is STANDARD-1 | NATIVE
NATIVE stands for computers own collating sequence whereas STANDARD-1 stands
for ASCII collating sequence.
IF/THEN/ELSE/END-IF
The most famous decision making statement in all language is IF. The syntax
of IF statement is given below: IF can be coded without any ELSE statement. THEN
is a noise word and it is optional.
If ORs & ANDs are used in the same sentence, ANDs are evaluated first from
left to right, followed by ORs. This rule can be overridden by using parentheses.
The permitted relation conditions are =, <, >, <=, >=, <>
CONTINUE is no operation statement. The control is just passed to next
STATEMENT. NEXT SENTENCE passes the control to the next SENTENCE. Referring to
first page, Sentence is defined as statement(s) ended with full-stop.
It is advised to use END-IF, explicit scope terminator for the IF statements
than period, implicit scope terminator.
IF condition1 AND condition2 THEN
Statement-Block-1
ELSE
IF condition3 THEN
CONTINUE
ELSE
IF condition4 THEN
Statement-Block-2
ELSE
NEXT SENTENCE
END-IF
END-IF
END-IF
Statement-Block-2 will be executed only when condition 1, 2 and 4 are TRUE
and condition 3 is FALSE.
Implied operand: In compound conditions, it is not always necessary to
specify both operands for each condition. IF TOTAL=7 or 8 is acceptable. Here
TOTAL=8 is implied operation.
SIGN test and CLASS test
SIGN test is used to check the sign of a data item. It can be done as follows
Page:60
IF identifier is POSITIVE/NEGATIVE/ZERO
CLASS test is used to check the content of data item against pre-defined range of
values. It can be done as follows IF identifier is NUMERIC/ALPHABETIC/ALPHABETIC-HIGHER/
ALPHABETIC-LOWER
We can define our own classes in the special names paragraph. We have defined a
class DIGIT in our special names paragraph. It can be used in the following way.
IF identifier is DIGIT
Negated conditions.
Any simple, relational, class, sign test can be negated using NOT.
But it is not always true that NOT NEGATIVE is equal to POSITIVE. (Example ZERO)
EVALUATE
With COBOL85, we use the EVALUATE verb to implement the case structure of
other languages. Multiple IF statements can be efficiently and effectively replaced
with EVALUATE statement. After the execution of one of the when clauses, the
control automatically comes to the statement following END-EVALUATE. Any complex
condition can be given in the WHEN clause. Break statement is not needed, as it is so
in other languages.
General Syntax
EVALUATE subject-1 (ALSO subject2..)
WHEN object-1 (ALSO object2..)
WHEN object-3 (ALSO object4..)
WHEN OTHER imperative statement
END-EVALUATE
1.Number of Subjects in EVALUATE clause should be equal to number of
objects in every WHEN clause.
2.Subject can be variable, expression or the keyword TRUE/ FLASE and
respectively objects can be values, TRUE/FALSE or any condition.
3.If none of the WHEN condition is satisfied, then WHEN OTHER path will be
executed.
Sample
EVALUATE SQLCODE ALSO TRUE
WHEN 100 ALSO A=B
WHEN -305 ALSO (A/C=4)
DISPLAY ALLOWED SQLCODE..PROCEEDING..
WHEN OTHER imperative statement
END-EVALUATE
In the above example, display will be thrown when one of the first two WHEN clauses
are true.
Page:61
PERFORM STATEMENTS
PERFORM will be useful when you want to execute a set of statements in
multiple places of the program. Write all the statements in one paragraph and invoke
it using PERFORM wherever needed. Once the paragraph is executed, the control
comes back to next statement following the PERFORM.
1.SIMPLE PERFORM.
PERFORM PARA-1.
DISPLAY PARA-1 executed
STOP RUN.
PARA-1.
Statement1
Statement2.
It executes all the instructions coded in PARA-1 and then transfers the control
to the next instruction in sequence.
2.INLINE PERFORM.
When sets of statements are used only in one place then we can group all of
them within PERFORM END-PERFORM structure. This is called INLINE PERFORM.
This is equal to DO..END structure of other languages.
PERFORM
ADD A TO B
MULTIPLE B BY C
DISPLAY VALUE OF A+B*C C
END-PERFORM
3. PERFORM PARA-1 THRU PARA-N.
All the paragraphs between PARA-1 and PARA-N are executed once.
4. PERFORM PARA-1 THRU PARA-N UNTIL condition(s).
The identifiers used in the UNTIL condition(s) must be altered within the
paragraph(s) being performed; otherwise the paragraphs will be performed
indefinitely. If the condition in the UNTIL clause is met at first time of execution, then
named paragraph(s) will not be executed at all.
5. PERFORM PARA-1 THRU PARA-N N TIMES.
N can be literal defined as numeric item in working storage or hard coded
constant.
6. PERFORM PARA-1 THRU PARA-N VARYING identifier1
FROM identifier 2 BY identifier3 UNTIL condition(s)
Initialize identifier1 with identifier2 and test the condition(s). If the condition
is false execute the statements in PARA-1 thru PARA-N and increment identifier1 BY
Page:62
identifier3 and check the condition(s) again. If the condition is again false, repeat
this process till the condition is satisfied.
7.PERFORM PARA-1 WITH TEST BEFORE/AFTER UNTIL condition(s).
With TEST BEFORE, Condition is checked first and if it found false, then PARA1 is executed and this is the default. (Functions like DO- WHILE)
With TEST AFTER, PARA-1 is executed once and then the condition is checked.
(Functions like DO-UNTIL)
Refer Table session for eighth type of PERFORM.
EXIT statement.
COBOL reserved word that performs NOTHING. It is used as a single
statement in a paragraph that indicate the end of paragraph(s) execution.
EXIT must be the only statement in a paragraph in COBOL74 whereas it can be used
with other statements in COBOL85.
GO TO Usage:
In a structured top-down programming GO TO is not preferable. It offers
permanent control transfer to another paragraph and the chances of logic errors is
much greater with GO TO than PERFORM. The readability of the program will also be
badly affected.
But still GO TO can be used within the paragraphs being performed. i.e. When
using the THRU option of PERFORM statement, branches or GO TO statements, are
permitted as long as they are within the range of named paragraphs.
PERFORM 100-STEP1 THRU STEP-4
..
100-STEP-1.
ADD A TO B GIVING C.
IF D = ZERO DISPLAY MULTIPLICATION NOT DONE
GO TO 300-STEP3
END-IF.
200-STEP-2.
MULTIPLY C BY D.
300-STEP-3.
DISPLAY VALUE OF C: C.
Here GO TO used within the range of PERFORM. This kind of Controlled GO TO is fine
with structured programming also!
Page:63
TABLES
An OCCURS clause is used to indicate the repeated occurrences of items of
the same format in a structure. OCCURS clause is not valid for 01, 77, 88 levels.
It can be defined as elementary or group item. Initialization of large table
occurrences with specific values are usually done using perform loops in procedure
division. Simple tables can be initialized in the following way.
01 WEEK-ARRAY VALUE MONTUEWEDTHUFRISATSUN.
05 WS-WEEK-DAYS OCCURS 7 TIMES PIC X(03).
Dynamic array is the array whose size is decided during runtime just before the
access of first element of the array.
01 WS-MONTH-DAY-CAL.
05 WS-DAYS OCCURS 31 TIMES DEPENDING ON WS-OCCURENCE.
IF MONTH = FEB MOVE 28 to WS-OCCURRENCE.
Array Items can be accessed using INDEX or subscript and the difference
between them are listed in the table. Relative subscripts and relative indexes are
supported only in COBOL85. Literals used in relative subscripting/indexing must be
an unsigned integer.
ADD WS-SAL(SUB) WS-SAL(SUB + 1) TO WS-SAL(SUB + 2).
Sl #
1
2
3
4
5
Subscript
Working Storage item
It means occurrence
Occurrence, in turn translated to
displacement to access elements
and so slower than INDEX access.
It can be used in any arithmetic
operations or for display.
Subscripts can be modified by any
arithmetic statement.
Index
Internal Item No need to declare it.
It means displacement
Faster and efficient.
It cannot be used for arithmetic
operation or for display purpose.
INDEX can only be modified with SET,
SEARCH and PERFORM statements.
Page:64
SEARCH
When the requirement is to randomly access sequential information, the only
possible way is to load the information in an array and look up in the array for
the information requested.
This table look up can be done in two ways:
1. Sequential Search (SEARCH)
2. Binary Search (SEARCH ALL)
To use SEARCH/SEARCH ALL, table should have an index. To use SEARCH ALL the
table should be in a sorted order.
Page:65
Difference between SEARCH and SEARCH ALL

Sl #
1
2
3
Sequential SEARCH
SEARCH
Table should have INDEX
Table need not be in SORTED
order.
Multiple WHEN conditions can be

coded.
Any logical comparison is possible.
5.
6
To start the search from first

element, index should be set to 1.
The logic is search one after other
from the current index till the
search is successful or end of
table.
Preferred in the following cases:

1. Table Size is small.
2. Table contains duplicates and
required to retrieve all the
duplicates.
3.Table can be loaded in such a
way that the most frequently
accessing elements can be loaded
in the first few occurrences.
Syntax:
SET indexname-1 TO 1.
SEARCH identifier-1 AT END
Display No Match:
WHEN condition-1
Statements /NEXT SENTENCE
WHEN condition-2
END-SEARCH
Identifier-1 should be OCCURS item
and not 01 item.
Binary SEARCH
SEARCH ALL
Table should have INDEX
Table should be in sorted order of the
searching argument. There should be
ASCENDING/DESCENDING Clause.
Only one WHEN condition can be coded.
Only = is possible. Only AND is possible
in compound conditions.
Index need not be set to 1. The logic is
compare the item to be searched with
the item at the mid of the table. If it
matches fine, else repeat the process
with left or right half depending on where
the item lies. For this logic to work fine,
the array should be sorted on the item
being searched.
Preferred in the following cases:
When the size of the table is large and all
the elements in the table will be
frequently accessed and table does not
contain any duplicates and contain a
unique key.
Syntax:
SEARCH ALL Identifier-1 AT END
Display No Match:
WHEN condition-1
END-SEARCH
AT END Clause and SET command:

Though AT END Clause is optional, it is highly recommended to code that. Because if
it is not coded and element looking for is not found, then the control simply comes to
the next statement after SEARCH where an invalid table item can be referred and
that may lead to incorrect results / abnormal ends.
SET statement Syntax:
SET index-name-1 TO/UP BY/DOWN BY integer-1.
Page:66
Multi Dimensional Arrays

COBOL74 supports array of maximum of three dimensions whereas COBOL85
supports up to seven dimensions. The lowest- level OCCURS data-name or an item
subordinate to it is used to access an entry in the array or the table.
If we use SEARCH for accessing multi-dimension table, then INDEXED BY
must be used on all OCCURS levels. Expanded nested perform is available for
processing multi level tables. The syntax of this perform is given below:
PERFORM para-1 thru para-n
VARYING index-1 from 1 BY 1 UNTIL index-1 > size- of- outer-occurs
AFTER VARYING index-2 from 1 by 1 until index-2 > size of inner occurs.
SEARCH example for multi level tables:
01 EMP-TABLE.
05 DEPTNUMBER OCCURS 10 TIMES INDEXED BY I1.
10 EMP-DETAIL OCCURS 50 TIMES INDEXED BY I2.
15 EMP-NUMBER PIC 9(04).
15 EMP-SALARY PIC 9(05).
77 EMPNUMBER-IN
PIC 9(04) VALUE 2052.
PERFORM 100-SEARCH-EMP-SAL VARYING I1 FROM 1 BY 1
UNTIL I1 > 10 OR WS-FOUND
100-SEARCH-EMP-SAL.
SET I2 TO 1.
SEARCH EMP-DETAIL AT END DISPLAY NOT FOUND
== > Lowest Occurs
WHEN EMPNUMBER-IN = EMP-NUMBER(I1,I2)
DISPLAY SALARY IS: EMP-SALARY(I1,I2)
SET WS-FOUND TO TRUE
== > Search ends
END-SEARCH.
NESTED PROGRAMS, GLOBAL, EXTERNAL
One program may contain other program(s). The contained program(s) may
themselves contain yet other program(s). All the contained and containing programs
should end with END PROGRAM statement. PGMB is nested a program in the
example below:
Example:
PROGRAM-ID. PGMA
PROGRAM-ID. PGMB
END PROGRAM PGMB.
END PROGRAM PGMA.

If you want access any working storage variable of PGMA in PGMB, then
declare them with the clause IS GLOBAL in PGMA. If you want to access any
working storage variable of PGMB in PGMA, declare them with the clause IS
EXTERNAL in PGMB. Nested Programs are supported only in COBOL85.
If there is a program PGMC inside PGMB, it cannot be called from PGMA
unless its program id is qualified with keyword COMMON.
Page:67
SORT and MERGE

The programming SORT is called as internal sort whereas the sort in JCL is
called external sort. If you want to manipulate the data before feeding to sort, prefer
internal sort. In all other cases, external sort is the good choice. Internal sort, in turn
invokes the SORT product of your installation. (DFSORT). In the run JCL, allocate at
least three sort work files. (SORT-WKnn => nn can be 00-99).
FASTSRT compiler option makes the DFSORT to do all file I-O operation than
your COBOL program. It would significantly improve the performance. The result of
the SORT can be checked in SORT-RETURN register. If the sort is successful, the
value will be 0 else 16.
Syntax:
SORT SORTFILE ON ASCENDING /DESCENDING KEY sd-key-1 sd-key2
USING file1 file2 / INPUT PROCEDURE IS section-1
GIVING file3
/ OUTPUT PROCEDURE is section-2
END-SORT
File1, File2 are to-be-sorted input files and File3 is sorted-output file and all
of them are defined in FD.SORTFILE is Disk SORT Work file that is defined at SD. It
should not be explicitly opened or closed.
INPUT PROCEDURE and USING are mutually exclusive. If USING is used, then
file1 and file2 should not be opened or READ explicitly. If INPUT PROCEDURE is used
then File1 and file2 need to be OPENed and READ the records one by one until end of
the file and pass the required records to sort-work-file using the command RELEASE.
Syntax: RELEASE sort-work-record from input-file-record.
OUTPUT PROCEDURE and GIVING are mutually exclusive. If GIVING is used,
then file3 should not be opened or WRITE explicitly. If OUTPUT procedure is used,
then File3 should be OPENed and the required records from sort work file should be
RETURNed to it. Once AT END is reached for sort-work-file, close the output file.
Syntax: RETURN sort-work-file-name AT END imperative statement.
INPUT PROCEDURE Vs OUTPUT PROCEDURE:
Sometimes it would be more efficient to process data before it is sorted,
whereas other times it is more efficient to process after it is sorted. If we intend to
eliminate more records, then it would be better preprocess them before feeding to
SORT. If we want to eliminate all the records having spaces in the key field then it
would be efficient if we eliminate them after sorting. Because the records with blank
key comes first after sorting.
MERGE
It is same as sort. USING is mandatory. There should be minimum two files in

USING.
MERGE Sort-work-file ON ASCENDING KEY dataname1 dataname2
USING file1 file2
GIVING file3 / OUTPUT PROCEDURE is section-1
END-MERGE
Program sort registers (and its equivalent DFSORT parameter/meaning)
SORT-FILE-SIZE (FILSZ), SORT-CORE-SIZE (RESINV), SORT-MSG(MSGDDN)
SORT-MODE-SIZE (SMS=nnnnn)
SORT-RETURN(return-code of sort) and
SORT-CONTROL (Names the file of control card default is IGZSRTCD)
Page:68
Page:69
STRING MANIPULATION
A string refers to a sequence of characters. String manipulation operations
include finding a particular character/sub-string in a string, replacing particular
character/sub-string in a string, concatenating strings and segmenting strings.
All these functions are handled by three verbs INSPECT, STRING and UNSTRING in
COBOL. EXAMINE is the obsolete version of INSPECT supported in COBOL74.
INSPECT- FOR COUNTING
It is used to tally the occurrence of a single character or groups of characters in a
data field.
INSPECT identifier-1 TALLYING identifier-2 FOR
ALL|LEADING literal-1|identifier-3
[BEFORE|AFTER INITIAL identifier-4|literal-2]
- Optional.
INSPECT identifier-1 TALLYING identifier-2 FOR

CHARACTERS
- Optional.
Main String is identifier-1 and count is stored in identifier-2. Literal-1 or

Identifier-3 is a character or group-of-characters you are looking in the main-string.
INSPECT further qualifies the search with BEFORE and AFTER of the initial occurrence
of identifier-4 or literal-2.
Example:
WS-NAME MUTHU SARAVANA SURYA CHANDRA DEVI
INSPECT WS-NAME TALLYING WS-COUNT ALL S
BEFORE INITIAL CHANDRA AFTER INITIAL SARAVANA
END-INSPECT
Result: WS-COUNT contains 1
INSPECT- FOR REPLACING
It is used to replace the occurrence of a single character or groups of
characters in a data field.
INSPECT identifier-1 REPLACING
ALL|LEADING literal-1|identifier-2 BY identifier-3|literal-2
- Optional.
INSPECT identifier-1 REPLACING CHARACTERS
BY identifier-2 BEFORE|AFTER INITIAL identifier-3|literal-1
INSPECT-FOR COUNTING AND REPLACING
It is a combination of the above two methods.
INSPECT identifier-1 TALLYING <tallying part > REPLACING <replacing part>
Page:70
STRING
STRING command is used to concatenate one or more strings.
Syntax:
STRING identifier-1 / literal-1, identifier-2/ literal-2
DELIMITED BY (identifier-3/literal-3/SIZE)
INTO identifier-4
END-STRING.
01 VAR1 PIC X(10) VALUE MUTHU
01 VAR2 PIC X(10) VALUE SARA
01 VAR2 PIC X(20).
To get display MUTHU,SARA

STRING VAR1 DELIMITED BY
, DELIMITED BY SIZE
VAR2 DELIMITED BY
INTO VAR3
END-STRING.
The receiving field must be an elementary data item with no editing symbols
and JUST RIGHT clause.
With STRING statement, specific characters of a string can be replaced
whereas MOVE replaces the full string.
01 AGE-OUT PIC X(12) VALUE 12 YEARS OLD.
STRING 18 DELIMITED BY SIZE INTO AGE-OUT. => 18 YEARS OLD.
Reference Modification equivalent of SUBSTR
Reference modification is used to retrieve or overwrite a sub-string of a
string. : is known as reference modification operator.
Syntax: String(Starting-Position:Length)
MOVE 18 TO AGE-OUT(1:2) does the same as what we did with STRING command.
When it is used in array elements, the syntax is
Array-element (occurrence) (Starting-Position:Length)
UNSTRING
UNSTRING command is used to split one string to many strings.
Syntax:
UNSTRING identifier-1
[DELIMITED BY (ALL/) identifier2/literal1 [,OR (ALL/) (identifier-3/literal-2),..]]
INTO identifier-4 [,DELIMITER IN identifier-5, COUNT IN identifier-6]
[,identifier-7 [,DELIMITER IN identifier-8, COUNT IN identifier-9]
01 WS-DATA PIC X(12) VALUE 10/200/300/1.
UNSTRING WS-DATA DELIMITED BY /
INTO WS-FLD1 DELIMITER IN WS-D1 COUNT IN WS-C1
WS-FLD2 DELIMITER IN WS-D2 COUNT IN WS-C2
WS-FLD3 DELIMITER IN WS-D3 COUNT IN WS-C3
END-UNSTRING.
Result:
WS-FLD1 = 10 WS-FLD2 =200 WS-FLD3=300
Page:71
WS-C1 = 2 WS-C2=3 WS-C3=3 WS-D1 = / WS-D2=/ WS-D3 /

ON OVERFLOW can be coded with STRING and UNSTRING. If there is STRING
truncation then the imperative statements followed ON OVERFLOW will be executed.
COPY Statement
A COPY statement is used to bring a series of prewritten COBOL entries that
have been stored in library, into a program.
1.Common routines like error routine, date validation routine are coded in a library
and bring into the program by COPY.
2. Master files are used in multiple programs. Their layout can be placed in one
copybook and be placed wherever the files are used. It promotes program
standardization since all the programs share the same layout and the same data
names.
This reduces coding and debugging time. Change in layout needs change in
copybook only. It is enough if we just recompile the program for making the new
copy effective.
Syntax:
COPY copybook-name [(OF/IN) library name]
[REPLACING string-to-be-replaced BY replacing-string]
Copybooks are stored as members in PDS library and during compilation time, they
are included into the program. By default, the copybook library is SYSLIB and it can
be changed using IN or OF of COPY statement.
Copybooks can be used in the following paragraphs.
SOURCE-COMPUTER, OBJECT-COMPUTER, SPECIAL-NAMES, FILE-CONTROL,
IO-CONTROL, FD SECTION, PARAGRAPHS IN PROCEDURE DIVISION.
If the same copybook is used more than once in the program, then there will
be duplicate data declaration error during compilation, as all the fields are declared
twice. In this case, one copybook can be used with REPLACING verb to replace highlevel qualifier of the all the variables with another qualifier.
Example:
COPY CUSTOMER REPLACING CUST1- BY CUST2-.
Delimiter = = should be used for replacing pseudo texts. The replacing

option does not alter the prewritten entries in the library; the changes are made to
the users source program only.
CALL statement (Sub-Programs)
When a specific functionality need to be performed in more than one
program, it is best to write them separately and call them into each program. Sub
Programs can be written in any programming language. They are typically written in
a language best suited to the specific task required and thus provide greater
flexibility.
Main Program Changes:
CALL statement is used for executing the sub-program from the main
program. A sample of CALL statement is given below:
Page:72
CALL PGM2 USING BY REFERENCE WS-VAR1, BY CONTENT WS-VAR2.

PGM2 is called here. WS-VAR1 and WS-VAR2 are working storage items.
WS-VAR1 is passed by reference. WS-VAR2 is passed by Content. BY REFERENCE is
default in COBOL and need not be coded. BY CONTENT LENGTH phrase permits the
length of data item to be passed to a called program.
Sub-Program Changes:
WS-VAR1 and WS-VAR2 are working storage items of main program.
As we have already mentioned, the linkage section is used for accessing external
elements. As these working storage items are owned by main program, to access
them in the sub-program, we need to define them in the linkage section.
LINKAGE SECTION.
01 LNK-ITEMS.
05 LK-VAR1 PIC 9(04).
05 LK-VAR2 PIC 9(04).
In addition to define them in linkage section, the procedure division should be coded
with these data items for address-ability.
PROCEDURE DIVISION USING LK-VAR1,LK-VAR2
The name of the identifiers in the called and calling program need not be the
same (WS-VAR1 & LK-VAR1).
The last statement of your sub-program should be EXIT PROGRAM. This
returns the control back to main program. GOBACK can also be coded instead of
EXIT PROGRAM but not STOP RUN. EXIT PROGRAM should be the only statement in a
paragraph in COBOL74 whereas it can be coded along with other statements in a
paragraph in COBOL85.
PROGRAM-ID. <Program-name> IS INITIAL PROGRAM.
If IS INITIAL PROGRAM is coded along with program-id of sub program, then
the program will be in initial stage every time it is called (COBOL85 feature).
Alternatively CANCEL issued after CALL, will set the sub-program to initial state.
If the sub program is modified then it needs to be recompiled. The need for
main program recompilation is decided by the compiler option used for the main
program. If the DYNAM compiler is used, then there is no need to recompile the main
program. The modified subroutine will be in effect during the run. NODYNAM is
default that expects the main program recompilation.
Difference between Pass-by-reference and Pass-by-content
Sl #
1
2
3
4
Passl By Reference
CALL sub1 USING BY
REFERENCE WS-VAR1
It is default in COBOL. BY
REFERENCE is not needed.
Address of WS-VAR1 is passed
The sub-program modifications
Pass By Content
CALL sub1 USING BY CONTENT WS-VAR1
(BY CONTENT keyword is needed)
BY CONTENT key word is mandatory to
pass an element by value.
Value of WS-VAR1 is passed
The sub-program modifications on the
Page:73
on the passed elements are

visible in the main program.
passed elements are local to that

sub-program and not visible in the main
program.
Difference between Static Call and Dynamic Call

Sl #
1
STATIC Call
Identified by Call literal.
Ex: CALL PGM1.
Default Compiler option is

NODYNAM and so all the literal
calls are considered as static calls.
3.
If the subprogram undergoes

change, sub program and main
program need to be recompiled.
Sub modules are link edited with
main module.
Size of load module will be large
Fast
Less flexible.
4
5
6
7
DYNAMIC Call
Identified by Call variable and the
variable should be populated at run time.
01 WS-PGM PIC X(08).
Move PGM1 to WS-PGM
CALL WS-PGM
If you want convert the literal calls into
DYNAMIC, the program should be
compiled with DYNAM option.
By default, call variables and any unresolved calls are considered as dynamic.
If the subprogram undergoes change,
recompilation of subprogram is enough.
Sub modules are picked up during run
time from the load library.
Size of load module will be less.
Slow compared to Static call.
More flexible.
INTRINSIC FUNCTIONS:
LENGTH
MAX
MIN
NUMVAL
NUMVAL-C
Returns the length of the PIC clause. Used for finding length of group
item that spanned across multiple levels.
Returns the content of the argument that contains the maximum value
Returns the content of the argument that contains the minimum value
Returns the numeric value represented by an alphanumeric character
string specified in the argument.
Same as NUMVAL but currency and decimal points are ignored during
conversion.
Returns 21 Chars alphanumeric value YYYYMMDDHHMMSSnnnnnn
CURRENT
DATE
INTEGER OF DATE
INTEGER OF DAY
DATE OF INTEGER
DAY OF INTEGER
Returns
Returns
Returns
Returns
INTEGER equivalent of Gregorian date passed.

INTEGER equivalent of Julian date passed.
Gregorian date for the integer passed.
Julian date for the integer passed.
Note: FUNCTION INTEGER OF DATE (01-01-1601) returns 1.
Page:74
Page:75
FILE HANDLING
A data file is collection of relevant records and a record is collection of
relevant fields. The file handling in COBOL program involves five steps.
Steps in file-handing
1.Allocation: The files used in the program should be declared in FILECONTROL paragraph of environment division. The mapping with JCL DDNAME is done
here. The file is allocated to your program by this statement.
2.Definition. The layout of the file and its attributes are defined in the FILE
SECTION of DATA DIVISION.
3.Open: Dataset is connected/readied to your program using OPEN
statement. The mode of OPEN decides the operation allowed and the initial pointer in
the dataset. For example, EXTEND mode allows only write access and the pointer is
kept on the end of file to append.
4.Process: Process the file as per requirement, using the I-O statements
provided by COBOL. (READ, WRITE, REWRITE and DELETE)
5. Close: After the processing, close the file to disconnect it from the
program.
Allocation of file - SELECT Statement
(ENVIRONMENT-> INPUT-OUTPUT-> FILE-CONTROL)
SELECT [OPTIONAL] FILENAME ASSIGN to DDNAME
ORGANIZATION IS SEQUENTIAL/INDEXED/RELATIVE
ACCESS IS SEQUNETIAL/RANDOM/DYNAMIC
RECORD KEY IS FILE-KEY1
RELATIVE KEY IS WS-RRN
ALTERNARE RECORD KEY IS FILE-KEY2 WITH DUPLICATES
ALTERNARE RECORD KEY IS FILE-KEY3 WITHOUT DUPLICATES
FILE STATUS IS WS-FILE-STAT1
[,WS-FILE-STAT2]
=>ALL Files
=>ALL Files
=>ALL Files
=>KSDS
=>RRDS
=>KSDS with
=>AIX
=>ALL Files
=>VSAM Files
SELECT Statement- OPTIONAL Clause

This can be coded only for input files. If OPTIONAL is not coded, then the
input file is expected to present in JCL. If not, an execution error will occur.
If OPTIONAL is coded, then if the file is not mapped in JCL, it is considered as
empty file and the first read results end of file.
The file can also be dynamically allocated instead of static allocation in JCL.
SELECT Statement- ASSIGN TO
FILENAME is the logical name used inside the program and DDNAME is the
logical name in the JCL, mapped with physical dataset. DDNAME can be prefixed with
UT-S- to indicate QSAM file, UT-AS- to indicate ESDS file and with no prefix to
indicate KSDS/RRDS file.
JCL Step executing the program should have a dataset with DDNAME as label
Page:76
//DDNAME DD DSN=BPMAIN.EMPLOYEE.DATA,DISP=SHR
SELECT Statement-ORGANIZATION
It can be SEQUENTIAL (PS or VSAM ESDS), INDEXED (VSAM KSDS),
RELATIVE (VSAM RRDS). Default is Sequential.
SELECT Statement-ACCESS MODE
SEQUENTIAL.
It is default access mode and it is used to access the records ONLY in
sequential order. To read 100th record, first 99 records need to be read and skipped.
RANDOM.
Records can be randomly accessed in the program using the
primary/alternate key of indexed file organization or relative record number of
relative organization.100th record can directly be read after getting the address of the
record from the INDEX part for INDEXED files.100th record can directly be read for
RELATIVE files even without any index.
DYNAMIC.
It is mixed access mode where the file can be accessed in random as well as
sequential mode in the program.
Example: Reading the details of all the employees between 1000-2000. First
randomly access 1000th employee record, then read sequentially till 2000th employee
record. START and READ NEXT commands are used for this purpose in the procedure
division.
SELECT Statement-RECORD KEY IS
It is primary key of VSAM KSDS file. It should be unique and part of indexed
record structure.
SELECT Statement-ALTERNATE RECORD KEY IS
This phrase is used for KSDS files defined with AIX. Add the clause WITH
DUPLICATES if the AIX is defined with duplicates.
Referring to VSAM basics, every alternate index record has an associated
PATH and the path should be allocated in the JCL that invokes this program.
The DDNAME of the path should be DDNAME of the base cluster suffixed with
1 for the first alternate record clause, suffixed with n for nth ALTERNATE RECORD
KEY clause in SELECT clause.
SELECT Statement-FILE STATUS IS WS-FILE-STAT1,WS-FILE-STAT2
WS-FILE-STAT1 should be defined as PIC X(02) in working storage section.
After every file operation, the file status should be checked for allowable values.
WS-FILE-STAT2 can be coded for VSAM files to get the VSAM return code (2
bytes), VSAM function-code (1 byte) and VSAM feedback code (3 bytes).
This is a 6- byte field in working storage.
RESERVE Clause.
RESERVE clause [RESERVE integer AREA ] can be coded in the SELECT
statement. The number of buffers to be allocated for the file is coded here.
By default two buffers will be allocated if the clause is not coded. Since similar option
is available in JCL, this is not coded in program.
Page:77
RESERVE 1 AREA allocates one buffer, for the file in the SELECT statement.
Defining the file in FILE SECTION - FD

FD FILENAME
RECORDING MODE IS V/VB/F/FB
RECORD CONTAINS M CHARACTERS (TO N CHARACTERS)
BLOCK CONTAINS X CHARACTERS/RECORDS (TO Y CHARACTERS/RECORDS)
LABEL RECORDS ARE OMITTED/STANDARD
DATA RECORD IS FILE-RECORD.
01 FILE-RECORD PIC X(nnn).
FD-RECORD CONTAINS
It specifies the length of the record in terms of bytes. (It will be RECORD
contains m to n CHARACTERS for variable format files)
FD-BLOCK CONTAINS
It specifies the physical record size. It can be mentioned as number of logical
records OR number of characters, that is multiple of logical record length. It is
suggested to code BLOCK CONTAINS 0 RECORDS so that system will decide the
optimum size for the file based on the device used for storing the file. BLOCK
CONTAINS clause is treated as comments for VSAM files.
Advantage of Blocking:
1.I-O time is reduced as n numbers of records are read into main memory buffer
during an I-O.
2.Inter record gap is removed and the gap exist only between blocks. So memory
wastage due to IRG is avoided.
FD-RECORDING MODE IS
It can be F (FIXED) V(VARIABLE) FB(FIXED BLOCK) VB(VARIABLE BLOCKED)
Variable record file identification:
If there is no recording mode/record contains clause, it is still possible to
identify variable length records. If there is an OCCURS depending on clause or there
are multiple 01 levels and every 01 level is of different size, then the file would be of
variable length. Multiple 01 levels in File section is an example for implicit
redefinition.
FD-LABEL RECORDS Clause
As a general rule, LABEL RECORDS are STANDARD is coded for Disk and Tape
files, LABEL RECORDS ARE OMITTED is coded for printer files. In COBOL74, this
clause is a mandatory clause whereas COBOL85 made this as optional.
FD-DATA RECORD IS Clause
It is used to name the data record(s) of the file. More than one record can be
coded here.
Page:78
OPEN STATEMENT
Syntax:
OPEN OPENMODE FILENAME
OPENMODE can be INPUT OUTPUT I-O EXTEND
INPUT
- File can be used ONLY-FOR-READ purpose.
OUTPUT
- File can be used ONLY-FOR-WRITE purpose.
I-O
- File can be used FOR READ, WRITE and REWRITE purpose.
EXTEND
- File can be used FOR appending records using WRITE.
CLOSE statement.
The used files are closed using CLOSE statement. If you dont close the files,
the completion of the program closes all the files used in the program.
Syntax:
CLOSE FILENAME
OPEN and CLOSE for TAPE files - Advanced
If more than one file is stored in a reel of tape, it is called as multi-file
volume. When one file is stored in more than one reel of tape, it is called as
multi-volume label. One reel is known as one volume. When the end of one volume is
reached, automatically the next volume opens. So there is no special control is
needed for multi volume files.
OPEN INPUT file-1 [WITH NO REWIND | REVERSED]
OPEN OUTPUT file-2 [WITH NO REWIND]
CLOSE file-3 [{REEL|UNIT} [WITH NO REWIND| FOR REMOVAL]
CLOSE file-3 [WITH NO REWIND|LOCK]
UNIT and REEL are synonyms.
After opening a TAPE file, the file is positioned at its beginning. When opening
the file if the clause REVERSED is coded, then the file can be read in the REVERESE
direction. (Provided hardware supports this feature)
When you close the file, the tape is normally rewound. The NO REWIND
clause specifies that the TAPE should be left in its current position.
CLOSE statement with REEL option closes the current reel alone. So the next
READ will get the first record of next REEL. This will be useful when you want skip all
the records in the first reel after n number of records processing.
Since TAPE is sequential device, if you create multiple files in the same TAPE,
then before opening the second file, first file should be closed. At any point of time,
you can have only one file is active in the program. In addition to this, you have to
code MULTIPLE FILE clause in the I-O control paragraph of environment division.
MULTIPLE FILE TAPE CONTAINS
OUT-FILE1 POSITION 1
OUT-FILE3 POSITION 3.
The files OUT-FILE1 and OUT-FILE3 used in the program are part of a same
TAPE and they exist in first and third position in the tape. Alternatively, this
information can be passed from JCL using LABEL parameter.
Page:79
READ statement
READ statement is used to read the record from the file.
Syntax:
READ FILENAME [INTO ws-record] [KEY IS FILE-KEY1]
[AT END/INVALID KEY imperative statement1]
[NOT AT END/NOT INVALID KEY imperative statement2]
END-READ
If INTO clause is coded, then the file is directly read into working storage
section record. It is preferred as it avoids another move of file-section-record to
working-storage-record followed by simple READ. READ-INTO is not preferred for
variable size records where the length of the record being read is not known.
KEY IS clause is used while accessing a record randomly using
primary/alternate record key.
AT END and NOT AT END are used during sequential READ of the file.
INVALID KEY and NOT INVALID KEY are used during random read of the file.
Before accessing the file randomly, the key field should have a value before READ.
WRITE Statement
Write statement is used to write a new record in the file. If the file is opened
in EXTEND mode, the record will be appended. If the file is opened in OUTPUT mode,
the record will be added at the current position.
Syntax:
WRITE FILE-RECORD [FROM ws-record]

[INVALID KEY imperative statement1]
END-WRITE
FROM clause avoids the explicit move of working storage record to file section record
before WRITE.
REWRITE Statement
REWRITE is used to update an already read record. To update a record in a
file, the file should be opened in I-O mode.
Syntax:
REWRITE FILE-RECORD [FROM ws-record]
END-REWRITE
START Statement
START is used with dynamic access mode of indexed files. It establishes the
current location in the cluster for READ NEXT statement. START itself does not
retrieve any record.
Syntax:
START FILENAME
KEY is EQUAL TO/NOT LESS THAN/GREATER THAN key-name
Page:80
END-START.
DELETE Statement
DELETE is used to delete the most recently read record in the file. To delete a
record, the file should be opened in I-O mode.
Syntax:
DELETE FILENAME RECORD
END-DELETE
Reports FBA:
Reports contain the printing control character in the first byte. The record format
will be FBA and the LRECL will be 133 in the JCL. The program can define
printing control character and populate it manually or can define the layout with
132 bytes and by means of program supported WRITE verbs like
AFTER/BEFORE/ADVANCING. In the second case printing control character will
be automatically added at the time of compilation by the default ADV compiler
option.
Printing control character
(blank)
0
+
1
Eqlt COBOL statement

After advancing 1 line
After advancing 2 lines
After new page
File Error Handling

There are chances for failure of any file I-O processing. The failure of an I-O
operation can be accepted or cannot be tolerated. The severity of failure has to be
defined in the program design stage.
Let us assume that we dont have any error handling in our program. In this
case, for example, if you dont have a specific record in the file, the random read of
that record would immediately terminate the program with error record not found.
Error Handling Clauses Provided by COBOL.
The sudden termination can be avoided by handling this error, with INVALID
KEY clause of READ. Based on the importance of the record and business rule, we
can continue our program with next record or terminate the program properly.
AT END is another error handling clause provided by COBOL. But there is no way to
handle all such errors in this way.
Assign file-status and take the responsibility.
The second method is, assigning file-status to the file in the SELECT clause
and checks the file status after each and every I-O and ensures that the value of
status code is one of the allowable values. If it is not an allowable return code, then
abnormally end the program with error statements that would be easier to debug.
But we have to do this checking after each and every I-O operation.
This is MOST PREFERRED ERROR HANDLING METHOD in structured programming.
DECLARATIVES USE statement
Page:81
COBOL provides an option to group all the possible errors of specific

operation(s) in a place and that will be automatically invoked during the respective
operation(s) of any file. This avoids redundant code.
This is done in DECLARATIVE section of the procedure division. DECLARATIVE
should be the first section in the procedure division if coded.
PROCEDURE DIVISION.
DECLARATIVES.
USE-PROCEDURE SECTION.
USE AFTER EXCEPTION PROCEDURE ON INPUT
ERROR-PROCEDURE.
Check the file-status code for validity.
END-DECLARATIVES.
Whenever there is an error in the processing of ANY FILE opened in INPUT
mode, then the control comes to ERROR-PROCEDURE. The validity of error should be
checked in this paragraph and allow or restrict the process down, based on severity
of error code.
The complete syntax of USE statements is:
USE AFTER STANDARD ERROR|EXCEPTION PROCEDURE ON
INPUT|OUTPUT|I-O|EXTEND| file-1
If INPUT is coded, the following procedure will be executed for every
operation involved in any file that is opened in INPUT mode. OUTPUT, I-O and
EXTEND have the same meaning but the mode is different.
If file name (file-1) is coded in the USE statement, then all the input-output
operation of that specific file will be checked.
ERROR and EXCEPTION are synonyms.
The Procedure written in a DECLARATIVE section should not refer to any nondeclarative procedure written after the end procedure and vice-versa.
USE FOR DEBUGGING ALL PROCEDURE .
DEBUG -DECLARATIVES -PARA.
DISPLAY TRACING
DEBUG-NAME
END DECLARATIVES.
This declarative section of the program will be executed before the start of execution
of each paragraph defined in the program. i.e. "TRACING paraname" will be
displayed before the execution of each paragraph. DEBUG-NAME is the special
register that contains the name of the paragraph being executed.
I-O-CONTROL - SAME AREA AND SAME RECORD AREA

RESERVE clause of SELECT statement specifies the number of buffers to be
allocated for a file. SAME AREA allows more than one file to use the same buffer
area. This will be very useful when the program must work with a limited memory
space. But the problem is only one file should be open at a time if SAME AREA is
coded.
Syntax: SAME AREA FOR file-1 file-2 file-3.
If SAME RECORD AREA is coded, then the buffer is not shared but only the
record area is shared. So more than one file can be in open state. We should be
Page:82
careful while filling in the record area of the output file. This may destroy the record
read most recently.
Syntax: SAME RECORD AREA FOR file-1 file-2 file-3.
SAME SORT AREA allows more than one sort/merge work files to use the
same area. The sort work files are automatically allocated when file is opened and
de-allocated when file is closed. As the sort file is automatically opened and closed
during a SORT and two sort files cannot be opened at a time, this clause may not be
useful.
Syntax: SAME SORT|SORT-MERGE AREA for file-1 file-2.
File-1 or file-2 should be a SD file.
I-O CONTROL- RERUN Clause
RERUN ON rescue FOR EVERY integer RECORDS on file-1
This will cause checkpoint to be taken for every integer-1 records processing of
file-1. If the program ABENDED before the complete processing of the file-1, then
the program will restart from integer+1ST record instead of first record. The rescue
file details should be mentioned outside the program and it varies from installation to
installation.
ENTRY statement
ENTRY statement establishes an alternate ENTRY point in a COBOL called
sub-program. When a CALL statement naming the alternate entry point is executed
in a calling program, control is transferred to the next executable statement following
the entry statement. Except when a CALL statement refers to an entry name, the
ENTRY statements are ignored at run-time.
Page:83
FILE STATUS CODES

It is a two-byte working storage item. The first byte denotes the general
category whereas second byte denotes the particular type of error message under
that category.
0
0
2
4
5
7
1
0
4
2
1
2
3
4
3
5
7
8
9
4
1
2
3
4
7
8
9
9
1
2
3
6
7
Successful OPEN/READ/WRITE Operation

Successful completion
Duplicate key was detected which is allowed as per definition of AIX.
Length of record just READ didnt conform to the fixed length attributes for
the file.
Referenced Optional file is not present during OPEN. If open mode is I-O or
EXTEND, then file will be created.
Open or Close statement is executed with a phrase that implies a tape file (ex
NO REWIND) whereas the file is not in TAPE.
When AT END condition fails
Sequential READ is attempted on
1.after the end of file is reached
2.optional file that is not present.
Sequential READ was attempted for a relative file and RRN is larger than the
maximum that can be stored in the relative key data item.
When INDEX Key fails
Sequence error exists for sequentially accessed index file.
Attempt was made to write a record that would create a duplicate key.
Record not found.(for keyed random access)
Space not found for WRITE
Permanent Open error
Opening a non-optional file that was not present.
Open mode is not permitted.
Open issued for a file that was closed previously with lock
File attribute mismatch-Open failed.
Logic error in opening/closing/deleting
OPEN a opened file.
CLOSE attempted for not opened file.
IO statement before the current REWRITE/DELETE is not successful.
REWRITE attempt with invalid length
READ file which is not opened in INPUT or I-O mode
WRITE file which is not opened in I-O OUPUT or EXTEND mode
DELETE or REWRITE file which is not opened in I-O mode.
Implementation defined
VSAM password error
Logic error
VSAM resource unavailable
No DD statement specified for VSAM file.
File integrity verified for VSAM file.
Page:84
COBOL COMPILATION
SYSPRINT
(Compiler listing)
PARM
(Compiler
Options)
SYSIN
(Source)
IGYCRCTL
(COBOL COMPILER)
SYSLIB
(Copybook Library)
SYSLMOD
(Load Module)
SYSPRINT
(Link edit messages)
SYSLIN(Object Module)
PARM
(Link
edit Options)
IEWL
(Link Editor)
SYSLIB
(Subroutine Library)
COMPILATION JCL:
//SMSXL86B JOB ,'COMPILATION JCL', MSGCLASS=Q,MSGLEVEL=(1,1),CLASS=C
//COMPILE1 EXEC PGM=IGYCRCTL, PARM=XREF,APO,ADV,MAP,LIST),REGION=0M
//STEPLIB
DD DSN=SYS1.COB2LIB,DISP=SHR
//SYSIN
DD DSN=SMSXL86.TEST.COBOL(SAMPGM01),DISP=SHR
//SYSLIB
DD DSN=SMSXL86.COPYLIB,DISP=SHR
//SYSLIN
DD DSN=&&LOADSET, DCB=(RECFM=FB,LRECL=80,BLKSIZE=3200),
//
DISP=(NEW,PASS),UNIT=SYSDA,SPACE=(CYL,(5,10),RLSE),
//SYSUT1
DD UNIT=&SYSDA,SPACE=(CYL,(1,10)) => Code SYSUT2 to UT7
//LINKEDT1 EXEC PGM=IEWL,COND=(4,LT)
//SYSLIN
DD DSN=&&LOADSET, DISP=(OLD,DELETE)
//SYSLMOD DD DSN=&&GOSET(SAMPGM01),DISP=(NEW,PASS),UNIT=SYSDA
//
SPACE=(CYL,1,1,1))
//SYSLIB
DD DSN=SMSXL86.LOADLIB,DISP=SHR
//SYSUT1
DD UNIT=SYSDA,SPACE=(CYL,(1,10))
//*** EXECUTE THE PROGRAM ***
//EXECUTE1 EXEC PGM=*.LINKEDT1.SYSLMOD,COND=(4,LT),REGION=0M
//STEPLIB
DD DSN=SMSXL86.LOADLIB,DISP=SHR
//
DD DSN=SYS1.SCEERUN,DISP=SHR
//SYSOUT
DD SYSOUT=*
//SYSPRINT
Page:85
DD SYSOUT=*
Compiler Options
The default options that were set up when your compiler was installed are in
effect for your program unless you override them with other options. To check the
default compiler options of your installation, do a compile and check in the
compilation listing.
Ways of overriding the default options
1.Compiler options can be passed to COBOL Compiler Program (IGYCRCTL) through
the PARM in JCL.
2.PROCESS or CBL statement with compiler options, can be placed before the
identification division.
3.If the organization uses any third party product or its own utility then these options
can be coded in the pre-defined line of the utility panel.
Precedence of Compiler Options
1. (Highest precedence). Installation defaults, fixed by the installation.
2. Options coded on PROCESS /CBL statement
3. Options coded on JCL PARM parameters
4. (Lowest Precedence). Installation defaults, but not fixed.
The complete list of compiler option is in the table:
Aspect
Source Language
Date Processing
Maps and Listing
Object Deck
generation
Object Code
Control
Debugging
Other
Compiler Option
APOST, CMPR2, CURRENCY, DBCS, LIB, NUMBER,
QUOTE, SEQUENCE, WORD
DATEPROC, INTDATE, YEARWINDOW
LANGUAGE, LINECOUNT, LIST, MAP, OFFSET, SOURCE, SPACE,
TERMINAL, VBREF, XREF
COMPILE, DECK, NAME, OBJECT, PGMNAME
ADV, AWO, DLL, EXPORTALL, FASTSRT, OPTIMIZE, NUMPROC,
OUTDD, TRUNC, ZWB
DUMP, FLAG, FLAGMIG, FLAGSTD, SSRANGE, TYPECHK
ADATA, ANALYZE, EXIT, IDLGEN
ADV: It is meaningful if your program has any printer files with WRITE..ADVANCING
keyword. The compiler adds one byte prefix to the original LRECL of printer files for
printing control purpose. If you are manually populating printing control character in
the program, then you can compile your program with NOADV.
DYNAM: Use DYNAM to cause separately compiled programs invoked through the
CALL literal statement to be loaded dynamically at run time. DYNAM causes dynamic
loads (for CALL) and deletes (for CANCEL) of separately compiled programs at object
time. Any CALL identifier statements that cannot be resolved in your program are
also treated as dynamic calls. When you specify DYNAM, RESIDENT is also put into
effect.
Page:86
LIST/OFFSET: LIST and OFFSET are mutually exclusive. If you use both, LIST will be
ignored. LIST is used to produce listing a listing of the assembler language expansion
of your code. OFFSET is used to produce a condensed Procedure Division listing.
With OFFSET, the procedure portion of the listing will contain line numbers,
statement references, and the location of the first instruction generated for each
statement. These options are useful for solving system ABENDS. Refer JCL session
for more details.
MAP:
Use MAP to produce a listing of the items you defined in the Data Division.
SSRANGE: If the program is compiled with SSRANGE option, then any attempt to
refer an area outside the region of the table will abnormally terminate with
protection exception, usually S0C4.It also avoids any meaningless operation on
reference modification like negative number in the starting position of reference
modification expression. If the program is compiled with NOSSRANGE, then the
program may proceed further with junk or irrelevant data. So usually the programs
are compiled with SSRANGE during development and testing.
RENT: A program compiled as RENT is generated as a reentrant object module. CICS
programs should be compiled with RENT option to share the same copy of the
program by multiple transactions (Multithreading)
RESIDENT: Use the RESIDENT option to request the COBOL Library Management
Feature. (The COBOL Library Management Feature causes most COBOL library
routines to be located dynamically at run time, instead of being link-edited with the
COBOL program.).CICS Programs should be compiled with RESIENT option.
XREF: Use XREF to get a sorted cross-reference listing. EBCDIC data-names and
procedure-names will be listed in alphanumeric order. It also includes listing, where
all the data-names that are referenced within your program and the line number
where they are defined. This is useful for identifying the fields that are defined but
not used anywhere after the development of new program.
Page:87
Difference between OS/VS COBOL and VS COBOL 2?

OS VS COBOL (COBOL74)
24 Bit addressing
Nested Programs are not allowed
Report write feature is dropped. (RW)

NOTES, ON statement, REMARKS are
removed.
VS COBOL2 (COBOL85)
Allows 31 bit addressing and makes
easier to develop large applications.
Nested Program are allowed Improves
productivity because different
programmers can develop them at the
same time. Promote data sharing and
data protection.
Explicit scope terminators are introduced.
POINTER clause is added.
INSPECT replaces EXAMINE with advance
capabilities.
Page:88
Virtual Storage Access Method

VSAM
Page:89
VSAM (Virtual Storage Access Method)

History
Access Method is an interface between the application program and physical
operation of storage devices. It is a component of operating system. VSAM is the first
access method that efficiently uses the virtual storage of MVS. It can manipulate only
the data that resides on a DASD. (Direct access storage device)
IBM introduced VSAM in 1973 as a replacement for several existing access
methods designed by it earlier.
KSDS (Key Sequenced Data Set) replaced ISAM (Indexed Sequential Access Method)
RRDS (Relative Record Data Set) replaced BDAM (Basic Direct Access Method)
ESDS (Entry Sequence Data Set) provide same function as normal sequential QSAM.
(Queued Sequential Access Method).
Initially VSAM had only ESDS and KSDS. RRDS and Alternate Index to KSDS
are introduced in 1979. DF/ EF (Data Facility extended Function) VSAM was
introduced in 1979 with Integrated Catalog Facility (ICF) to replace the old VSAM
catalog of the previous versions.
The latest version of DFP/ VSAM released in 1991 called DFP/VSAM 3.3
contains enhancements like variable record length support for RRDS and added
DFSMS facilities.
Advantages of VSAM over other access methods
1.Data retrieval will be faster because of an efficiently organized index. The index is
small because it uses a key compression algorithm.
2.Insertion of records is easy due to embedded free space in the cluster.
3.Records can be physically deleted and the spaces used by them can be used for
storing other records without reorganization.
4.VSAM datasets can be shared across the regions and systems.
5.Datasets can be physically distributed over various volumes based on key ranges.
6.VSAM is independent of storage device types.
7.Information about VSAM datasets is centrally stored in VSAM catalog.
So referencing of any VSAM dataset need not be detailed in JCL.
Disadvantages of VSAM
1.To allow easy manipulation of records, free space should be left in the dataset and
this increases the spaces required.
2.Integrity of the dataset across the region and system need to be controlled by
user.
Page:90
CLUSTER
A cluster can be thought of as a logical dataset consisting of two separate
physical datasets:
1 The data component (contains the actual data).
2 The index component (contains the actual index).
All types of VSAM datasets are called clusters even though KSDS is the only
type that fulfills the cluster concept. ESDS and RRDS dont have Index component.
Data Component.
Control Intervals and Control Areas
VSAM stores records in the data component of the Cluster in units called
control intervals.
The control interval is the VSAM equivalent for a block and it is the unit of
data that is actually transmitted when records are read or written. Thus many logical
records form a control interval and many control intervals form a control area.
The Control Area (CA) is a fixed length unit of contiguous DASD storage
containing a number of CI s. The control area is VSAM internal unit for allocating
space within a cluster. The primary and secondary allocations consist of some
number of CA. Minimum size of control area is 1 track and maximum size is 1
cylinder.
Format of Control Interval
There are four separate areas within a control Interval.
1. Logical Record Area (LRA) that contains the records.
2. Free Space (FSPC). This area can be used for adding new records.
3. Unused Space. As FSPC may not be a multiple of record length, there will be
always some unused space in every CI. This can be minimized for fixed length
records by selecting proper control interval size.
4. Control Fields.
CIDF Control Interval Definition Field - 4 bytes field containing information
on free space availability in the control interval. One per control interval.
RDF Record Definition Field 3 bytes field. For fixed length records, there
will be 2 RDF, first contains the number of records in the control interval and
the second contains record length. For variable length records, the number of
RDF can vary depending on how many adjacent records have the same length
in the CI. If no two adjacent records are of the same length, then one RDF is
needed to describe each record.
Index Component (Sequence Set and Index Set)
Besides the data component, VSAM creates an index component. Index
component consists of index set and sequence set. The sequence set is the lowest
level of the index is called the sequence set and it contains primary keys and
pointers to the control intervals of the data component.
There is one sequence set for one control area. The highest record key of
every control interval is stored in sequence set. The highest record key of the
sequence set is stored in first level of index set. Based on the size of control interval
of index component, there will be 1-3 levels of index sets in the index component of
the dataset.
Page:91
Control Interval and Control Area Split

When a VSAM dataset is created, the free space parameter defined can
specify a percentage of control intervals to be left free during the load of VSAM file.
Later, when you add a record to the VSAM file, according to the key sequence, it is
placed in a specific control interval.
But if the specific control interval is already full, then this cannot be placed in
the sequence. The result is Control Interval split. It moves half of the records in the
filled control interval to any other free control interval available in the control area
and makes room for the new record in the right place.
If there is no free control interval exist in the control area, then half the
control intervals are moved to new control area and this is called control area split.
Example
1. In the example below, there are four control areas and every control area contains
two control intervals. Control fields are not shown in the diagram. There should be
one sequence for every control area. So there are four sequence sets. There are two
levels of index set. The second level of index set contains pointers to sequence set.
2. Control Interval Split: When a record with key 22 is added in the program, it
should physically be stored between the existing records 21 and 23. 21 and 23 are in
the first control interval of control area-3. There is no more space available to store
this record. So control split will occurs. Record 20 and 21 continue to exist in the
current control interval. Records 22 and 23 will be moved to any of the free control
intervals. In our case control interval 2 is free. So they are moved there and index
set is accordingly updated.
3. Control Area Split: When a record with key 3 in the program, it should be placed
between 2 and 4. These records are in first control interval of first control area and
there is no free space. So control interval split is expected. But there is no free
control interval in the control area-1. So control area split occur. New control area is
allocated and half the records of control area-1 will be moved there and indexes are
properly updated.
How index could make the access faster?
I faced an interesting question in an interview. The question follows:
My sequential file has 50 records. To get the 50th record, I have to read 49
data records and bypass. Indexes store primary keys with actual location. So to read
50th record using indexed organization, I have to get the location of 50th record from
index. So I have to read and bypass the location of 49 records. The only difference is
in the second case is I am doing sequential read in index set instead of dataset. How
do you say my index access would be faster?
To answer this question, you should know how indexes are organized and
read. We have already seen how they are organized. In the example below, to read
the 50th record, first root index is read and identified that the record should be in the
right hand side. In the second I-O, I will get the sequence and in third I-O, I will get
the location of the record. I will get my record in the fourth I-O instead 51 I-O. (50
IO in index and 1 I-O for getting the data)
Page:92
If I am accessing the first record, then sequential read needs only one I-O but
obviously my random read needs more. So we prefer indexed organization only when
the number of records is significant.
* 17
*4
*8
*17
*8
*14
12 14
17
CA-1
CA-2
Root Index
*50
* 23
*17
*23
20 21 23
First Level
Index
*50
*35
*50
33 35
Sequence set
Data Component
50
CA-3
CA-4
Properties of VSAM datasets

Properties of ESDS, KSDS and RRDS are listed a tabular format in the next
page. Though the table says RRDS records should be of fixed length, MVS/DFP VSAM
Version 3.3 allows variable length RRDS. VSAM internally implements it using KSDS.
VSAM LDS Properties: Linear datasets have no records. They are just long strings of
bytes. Since it is a long string of bytes an LDS has no FSPC, US, CIDF and RDF. The
CISZ is always 4k bytes. The main use of LDS is to implement other data access
organizations especially a relational database system like DB2.
As the data in an LDS is manipulated the operating system automatically
pages in and out the portions of the dataset being worked on. The dataset is
addressed by the RBA as if it were in memory and the system pages the needed
pages in and out. Thus the process is very simple and fast.
ESDS differs from QSAM file in the following ways:
1. QSAM can be created over TAPE whereas ESDS cannot reside on TAPE.
2. Alternate Index can be created over ESDS. This alternate index has alternate
key mapped with RBA. So Indexed and Random access are possible using the
alternate Index in ESDS whereas it is not possible with QSAM.
Page:93
Comparison of ESDS, KSDS and RRDS

Characteristics
Entry Sequence
ESDS
Based on entry
sequence
Access
Only Sequential
access is possible.
Alternate INDEX
May have one or more

alternate indexes. But
cannot be used in
BATCH COBOL. Can be
used in CICS COBOL.
A record RBA cannot
be changed
Location of the
record
Free Space
Exist at the end of the

dataset to add
records.
Deletion
Cannot be deleted.
REWRITE of same
length is possible.
Fixed or Variable
Possible
Record Size
SPANNED
records
Speciality
Preference
Occupies less Space

Application that
require sequential
access only.
Ex: PAYROLL
PROCESSING.
KSDS
Based on collating
sequence by key
field
Sequential and
random access is
possible. Random
access is thru
primary/alternate
key.
May have one or
more alternate
indexes.
RRDS
Based on relative
record number
order
Can be accessed
directly using
relative record
number, which
serves as address.
A record RBA can

be changed.
A relative record
number can be
changed.
Free slots are
available for adding
records at their
location.
Distributed free
space for inserting
records in between
or change the
length of existing
record.
DELETE is possible.
Not applicable
DELETE is possible.
Fixed or Variable
Possible
Fixed.
Not possible
Easy RANDOM
access and most
popular method
Applications that
require each record
to have a key field
and require both
direct and
sequential access.
Ex: Banking
application
Fastest access
method.
Applications that
require only direct
access. There
should be a field in
the record that can
be easily mapped
to RRN.
Page:94
IDC Access Method Services (IDCAMS)

AMS is used to perform various functions on VSAM datasets and catalogs.
AMS has got a utility program called IDCAMS. The functions of AMS are performed
using the different functional commands of IDCAMS. It can be invoked in the
following three ways:
1. Batch mode with JCL statements.
2. Interactively with TSO commands.
3. Calls from an application program.
Important functional commands of IDCAMS are:
1.
2.
3.
4.
5.
6.
7.
DEFINE
To create objects KSDS/ESDS/RRDS/GDG/VSAMSPACE etc
ALTER
- To alter the parameters of the object already exists.
PRINT
- To print and view the selected records of dataset.
DELETE
Delete the objects.
LISTCAT
View the complete information about any object.
REPRO
- Copy/Restore/Merge Utility for VSAM and NON-VSAM files.
EXPORT and IMPORT
Copy and restore datasets across the system.
8. VERIFY
To properly close the unclosed/ABENDED VSAM dataset.
Sample IDCAMS JCL
//JS10 EXEC PGM=IDCAMS,REGION=1024K, PARM=parameters
//STEPCAT DD DSN=..,DISP=SHR Optional STEPCAT
IDCAMS Messages
//SYSIN
DD *
Control statements
/*
//
Guidelines for coding commands
1. At least one blank space must be there between the command and the object and
between sub parameter values.
2. A space is optional between a parameter and the parenthesis enclosing its values.
3. Sub-parameter values can be separated using a space or a comma.
4. Multiple parameters can be coded on a line but it is better to place each on a line
by itself for better readability.
5. A parameter and a sub-parameter cannot be separated with a hyphen. A plus sign
should be used for this purpose.
Page:95
Meaning of FILE and DATASET in AMS commands

1. Keyword FILE should point to DDNAME and it is logical. There should be mapping
of the DDNAME with actual dataset. JES allocates the datasets associated with the
file just before the execution of step.
2. Keyword DATASET should directly point to the physical dataset and IDCAMS
allocates this file dynamically for operation.
DEFINE CLUSTER
This command is used to create and name a VSAM Cluster.
Basic Parameters for Define Cluster
DEFINE CLUSTER-NAME
This parameter specifies name to the VSAM cluster. The cluster name
becomes the dataset name in any JCL that invokes the cluster
// INPUT DD DSN= SMSXL86.TEST.VSAM, DISP=SHR
The name for a VSAM dataset can include up to 44 alphanumeric characters.
When the data and index parameters are coded to create the data and index
components, the name parameter is coded for them as well. If the name parameter
is omitted for the data and index VSAM tries to append part of .DATA or .INDEX as
appropriate as the low level qualifier depending on how many characters the dataset
name contains already and still staying within the 44 character limit.
If data and index components are named, parameter values can be applied
separately. This gives performance advantages for large datasets.
DEFINE CLUSTER-DATA COMPONENT
The data parameter instructs IDCAMS that a separate data component is to
be created. Data is optional but if coded must follow all parameters that apply to the
cluster as a whole. There are several options for the data parameter but name is the
most common.
DEFINE CLUSTER-INDEX COMPONENT
The index parameter creates a separate index component. Index is optional
but if coded must follow all of the parameters that apply only to the data component.
ESDS and RRDS must not have INDEX part.
DEFINE CLUSTER-SPACE Allocation
A VSAM dataset can have 123 extents in a VOLUME. Primary space is
allocated initially when the dataset is created and based on request secondary space
will be allocated. In the best case, you will get (primary +122 * secondary).
Refer the JCL section-SPACE parameter section for more details on EXTENTS.
VSAM calculates the control area size internally. Control area can of one
cylinder, the largest permitted by VSAM, usually yields the best performance. So it is
always better to allocate space in cylinders because this ensures a CA size of one
cylinder.
Page:96
The RECORDS parameter is used to allocate space in units of records for small
datasets. When this is done the RECORDSIZE parameter must be specified.
If allocation is specified in units of KILOBYTES or MEGABYTES VSAM reserves space
on the minimum number of tracks it needs to satisfy the request.
Syntax: UNIT(primary secondary)
UNIT can be CYL/CYLINDERS TRK/TRACKS REC/RECORDSD
KB/KILOBYTES MB/MEGABYTES
DEFINE CLUSTER-VOLUMES Parameter

The volumes parameter assigns one or more storage volumes to the dataset.
Multiple volumes can be specified and if so, they must be of the same device type.
The data and index components can be stored in separate volumes to provide a
performance advantage for large datasets.
VOLUMES(volser)
OR
VOLUMES(volser1 volser2)
DEFINE CLUSTER-RECORDSIZE Parameter (RECSZ)
The record size parameters specify VSAM what records to expect. The AVG
and MAX are the average and maximum values for variable length records. If records
are of fixed length AVG and MAX can be the same. Record size can be assigned at
the cluster or data level.
Syntax: RECORDSIZE(AVG MAX)
DEFINE CLUSTER- KEYS Parameter
The keys parameter defines the length and offset of the primary key in a
KSDS record. The offset is the primary keys displacement in bytes from the
beginning of the record. This parameter is defined for a KSDS only. Default is
KEYS(64 0).
Syntax: KEYS(length offset)
DEFINE CLUSTER- Dataset Type Parameter
The dataset type parameter specifies whether the dataset is
INDEXED(KSDS) , NONINDEXED(ESDS), NUMBERED(RRDS) OR LINEAR (LDS).
INDEXED (IXD) specifies a KSDS and it is the DEFAULT. When this parameter
is specified VSAM automatically creates and catalogs an index (if other parameters
like keys are valid). Indexed is also used for a variable length RRDS.
NONINDEXED(NIXD) when specified denotes an ESDS. No index is created
and records are accessed sequentially or by their relative byte address (RBA).
NUMBERED (NUMD) specifies an RRDS and LINEAR specifies an LDS.
Performance Parameters For DEFINE CLUSTER
The performance parameters are coded in the DEFINE CLUSTER command to
1. Reduce the amount of physical I-O.
2. Make necessary I-O faster.
3. Optimize disk utilization.
DEFINE CLUSTER- CONTROL INTERVAL SIZE Parameter
Page:97
The size of the CI is specified by the CISZ parameter. The size of the CI is
specified in bytes and it should be a multiple of 512 or 2048 depending on the type
of catalog (ICF or VSAM) and the length of the record.
For datasets cataloged in ICF catalogs the control interval should be a
multiple of 512 bytes with a range of 512 to 32768 bytes.
For datasets cataloged in VSAM catalogs, the data CISZ must be a multiple of
512 if records are of 8192 bytes or less, and a multiple of 2048 if records are of
more than 8192 bytes. If a CISZ, which is not a multiple of the above two is
assigned VSAM rounds the CISZ value up to the next highest multiple if necessary.
Best control interval size selection

For sequential processing of a KSDS, a relatively large CISZ will reduce
physical I/O by keeping more records accessible in the buffers. On the other hand,
for random processing of a KSDS a smaller CISZ would require lesser data transfer
time and fewer buffers, thus making I/O faster.
For an ESDS (since it is processed sequentially) the CISZ should be relatively large
depending on the size of the record.
Since an RRDS is processed randomly the CISZ should be relatively small.
The FREESPACE Parameter (FSPC)
The FREESPACE parameter is applicable to the KSDS and the variable-length
RRDS only. FREESPACE cannot be assigned to an ESDS because all new records are
added to the end of the dataset.
This parameter allocates some percentage of both the CI and CA for planned
free space, which can be used for adding new records or expanding existing variable
records. The parameter is coded as follows
FREESPACE(ci% ca%)
FREESPACE(ci% )
control interval only
FREESPACE(0 ca%)
control area only
FREESPACE(0 0) is the default.
FREESPACE(100 100) means only one record will be loaded in every control interval
and only one control interval will be loaded in every control area.
In order to effectively allocate FREESPACE the following factors have to be taken into
consideration.
1. The expected rate of growth: If even growth is expected apply FREESPACE to both
CI and CA. If uneven growth is expected apply FREESPACE only to the CA.
2. The expected number of records to be deleted.
3. How often the dataset will be reorganized with REPRO.
4. The performance requirements.
Page:98
The CI FREESPACE allocation should be enough to cover the length of one

record plus any additional RDFs that may result from a variable record length.
DEFINE CLUSTER- SPEED And RECOVERY Parameters
RECOVERY pre formats a VSAM dataset at the time of initial load or resume
load. It takes longer time to load the dataset through RECOVERY than through
SPEED which does not pre format the dataset. However if the load operation is
aborted in the middle, RECOVERY enables resumption immediately after the last
record loaded.
If SPEED is specified load will have to be restarted from the beginning. SPEED
is highly recommended because it will speed up the initial data load.
The BUFFERSPACE Parameter (BUFSP)

By default, VSAM allocates two data buffers for all types of datasets. One data
buffer for processing and reserves the second for potential split activity. In addition
to this, it would allocate one index buffer for a KSDS.
This parameter is used to override the default values. The BUFFERSPACE
parameter represents the amount of storage in bytes required to process the
contents of a minimum of one CI worth of data.
Syntax: BUFFERSPACE(bytes)
For sequential processing more number of data buffers need to be allocated.
For random processing more index buffers may be required, at least one for each
level of the index. For dynamic processing additional data and index buffers are
required in proportion to the ratio of sequential and random processing planned.
The BUFSP value provided gets translated into data buffers for an ESDS or
RRDS. But for a KSDS, VSAM decides on the number of data and index buffers based
on the access method specified in the application program.
If BUFSP is specified in the DEFINE CLUSTER command, all applications that
use the dataset can use only this allocation unless they override it in the BUFSP sub
parameter of the Access Method Parameter of JCL.
DEFINE CLUSTER-SPANNED Parameter (SPND)
The SPANNED parameter allows large records to span more than one CI.
If maximum record length is more than the CI size specified, allocation will fail unless
the SPANNED parameter is specified. However records cannot span across control
areas. The resulting free space in the spanned CI is unusable by other records even
if they fit logically in the unused bytes. NONSPANNED is the default.
It can be specified for ESDS and KSDS only.
DEFINE CLUSTER-The KEYRANGES Parameter (KYRNG)
The KEYRANGES parameter divides a large dataset among several volumes.
The volumes in which the records are to be placed are specified using the VOLUMES
parameter. Unless the KEYRANGES parameter is specified with the ORDERED
parameter, the records are assigned randomly to the volumes. These parameters are
illustrated by the following example
Page:99
DEFINE CLUSTER
(NAME(NTCI.V.UE4.W20000.T30.AV.DW200006)
CYL(5 1)
KEYS(8 0)
RECSZ(80 80)
KEYRANGES ((00000001 2999999)
(30000000 4700000)
(47000001 9999999))
VOLUMES
(NTTSOB
NTTSOJ
NTTSO5)
ORDERED
NOREUSE
INDEXED
_ _ _ more parameters
When the ORDERED parameter is coded the number of VOLUMES and KEYRANGES
must be the same.
DEFINE CLUSTER- REUSE Parameter (RUS)
This specifies that a cluster can be opened again and again as a reusable
cluster. It means, whenever you open the dataset in OUTPUT mode, all the records
that are already exist in the dataset are logically deleted. NOREUSE (UNRUS) is the
default and specifies the cluster as non-reusable.
A cluster cannot be reused if
1. KEYRANGES parameter is coded for it
2. An alternate index is built for it
3. The cluster has its own data space in both the VSAM and ICF catalog
environments.
DEFINE CLUSTER-IMBED And REPLICATE Parameters (IMBD/ REPL)
These parameters are applicable to a KSDS only. The IMBED parameter
implies that the sequence set (lowermost level) of the index component of a KSDS
will be placed on the first track of a data component CA and will be duplicated as
many times as possible on that track.
What IMBED does for a sequence set REPLICATE does for an index set. The
REPLICATE parameter forces each CI of the index set of the index component to be
written on a separate track and replicated as many times as possible. This parameter
reduces rotational delay when VSAM is accessing high-level index CI.
The IMBED option reduces the seek time it takes for the read-write head to
move from the index to the data component and the replication factor reduces the
rotational delay of the revolving disk.
NOIMBED(NIMBD) and NOREPLICATE(NREPL) are the defaults.
DEFINE CLUSTER-WRITECHECK Parameter (WCK)
This parameter instructs VSAM to invoke a DASD verify facility whenever a
record is written. NOWRITECHECK (NWCK) is the default and provides no DASD
verification. Since contemporary DASD devices are very reliable and because of the
high performance overhead associated with this parameter it is better to accept the
NOWRITECHECK option.
Page:100
DEFINE CLUSTER-SHAREOPTIONS Parameter

This parameter specifies how a VSAM dataset can be shared across the
regions and across the system. Default value is (1 3)
Syntax: SHAREOPTIONS(cross-region
Type
CR
Value
1
CR
CR
CS
CS
3
4
cross-system)
Meaning
READ and WRITE Integrity. Any number of jobs can read the
dataset OR only one job can write to the dataset.
ONLY WRITE Integrity. Any number of jobs can read the dataset
AND one job can write to the dataset.
NO Integrity. File is fully shared. It is programmer responsibility to
take proper lock over the file before use. Default value for CS.
Same as 3 but additionally forces a buffer refresh for each random
access.
DEFINE CLUSTER-ERASE Parameter

The ERASE parameter instructs VSAM to overwrite sensitive data with binary
zeroes when the cluster is deleted. NOERASE is the default and it means that the
deleted cluster is not to be overwritten with binary zeroes. ERASE can be coded at
the cluster or data level.
When applied to the cluster level it refers only to the data component since
the index component contains only keys and not data. ERASE can be specified with
the ALTER and DELETE commands. ERASE provided with DELETE overrides both
DEFINE and ALTER specifications.
Password Protection
VSAM datasets can be password protected at four different levels. Each level
gives a different access capability to the dataset. The levels are
1.
READPW- provides read only capability.
2.
UPDATEPW- records may be read, updated , added or deleted at this level.
3.
CONTROLPW- provides the programmer with the access capabilities of
READPW and UPDATEPW.
4.
MASTERPW- all the above operations plus the authority to delete the dataset
is provided.
Passwords provided at the cluster level protect only if access requires using
the clusters name as dataset name. Therefore it is advisable to protect the data and
index components using passwords because someone could otherwise access them
by name. Another feature of MVS called Resource Access Control Facility (RACF)
ignores VSAM passwords and imposes its own security and for most VSAM datasets
RACF security is sufficient.
The ATTEMPTS parameter coded with the password parameters specifies the
number of attempts permitted for the operator to enter the password before
abending the step.
Page:101
The CODE parameter allows for the specification of a code to display to the
operator in place of the entry name prompt.
The AUTHORIZATION parameter provides for additional security by naming an
assembler User Security Verification Routine (USVR). The sub parameter for this
enclosed in parenthesis is the entry point of the routine.
TO And FOR Parameters
When a dataset is allocated by Access Method Services the TO and FOR
parameters are two mutually exclusive parameters given to specify the retention
period of the cluster being defined.
TO(YYDDD)
or
FOR(nnnn)
YYDDD is Julian date & nnnn can be 1-9999
About CATALOG parameter

Most of the AMS commands provide CATALOG option. If the command is
DEFINE, then the dataset being defined will be placed in the catalog, mentioned in
the CATALOG parameter. Similarly while accessing the dataset,
1. The dataset is first searched in the catalog mentioned in CATALOG parameter.
2. If CATALOG is not coded, then the dataset is searched into the catalog coded in
STEPCAT or JOBCAT catalog.
3. If there is no STEPCAT and JOBCAT, then the dataset is searched in the user
catalog corresponding to the high level qualifier of the dataset.
4. If there is no user catalog found, then the dataset is looked into system catalog.
5. If the dataset is not listed in system catalog also, then you will get the error
message of dataset not found.
There will be one master catalog and n number of user catalogs in the
system. Every user catalog should have an entry in master catalog.
KSDS-ESDS-RRDS-LDS Sample Definitions
KSDS Definition
//KSDSMAKE EXEC PGM=IDCAMS
//SYSPRINT DD SYSOUT=A
//SYSIN DD*
DEFINE CLUSTER
(NAME(EMPLOYEE.KSDS.CLUSTER)
VOLUMES(VSAM02)
CYLINDERS(2,1)
CONTROL INTERVAL SIZE(4096)
FREESPACE(10,20)
KEYS(9,0)
RECORDSIZE(50,50))
DATA
(NAME(EMPLOYEE.KSDS.DATA))
INDEX
(NAME(EMPLOYEE.KSDS.INDEX))
CATALOG(VSAM.USERCAT.TWO)
/*
ESDS Definition
//ESDSMAKE EXEC PGM=IDCAMS
//SYSIN DD*
DEFINE CLUSTER
(NAME (EMPLOYEE.ESDS.CLUSTER)
VOLUMES (VSAM02)
CYLINDERS (2,1)
CONTROLINTERVALSIZE (4096)
RECORDSIZE (50,50)
NONINDEXED)
DATA
(NAME (EMPLOYEE.KSDS.DATA))
CATALOG (VSAM.USERCAT.TWO)
/*
Page:102
RRDS Definition
//RRDSMAKE EXEC PGM=IDCAMS
//SYSIN DD*
DEFINE CLUSTER
(NAME(EMPLOYEE.RRDS.CLUSTER)
VOLUMES(VSAM02)
CYLINDERS(2,1)
RECORDSIZE(50,50)
NUMBERED)
DATA
/*
LDS Definition
//LDSMAKE EXEC PGM=IDCAMS
//SYSIN DD*
DEFINE CLUSTER
(NAME(EMPLOYEE.RRDS.CLUSTER)
VOLUMES(VSAM02)
CYLINDERS(2,1)
LINEAR)
DATA
/*
IDCAMS - REPRO command

It is the general-purpose command that can operate on both VSAM and nonVSAM datasets. It performs three basic functions:
1. It loads an empty VSAM cluster with records. The data and index components for
KSDS are built automatically.
2. It creates a backup of a VSAM dataset on a physical sequential dataset and later
it can be used for restore/rebuild the VSAM dataset.
3. It merges data from two VSAM datasets.
Command Syntax
REPRO
INFILE(DDNAME) | INDATASET(DATASET-NAME)
OUTFILE(DDNAME) | OUTDATASET(DATASET-NAME)
optional parameters
INFILE/INDATASET and OUTFILE/OUTDATASET point to input and output datasets.

When loading a KSDS using REPRO, the input data should be first sorted in
ascending sequence by the field that will become the primary key in the output
dataset. When loading an ESDS the sort step can be eliminated since the records are
loaded in entry sequence. For an RRDS, the records are loaded in relative record
sequence starting with 1. The dataset should be sorted on the field that correlates to
the relative record number.
REPRO Record Selection
Page:103
REPRO can be used for selective copying.

Where to start REPRO
Where to stop REPRO
Valid For
FROMKEY(REC-KEY)
TOKEY(REC-KEY)
KSDS, ISAM
FROMADDRESS(RAB)
TOADDRESS(RBA)
KSDS, ESDS
FROMNUMBER(RRN)
TONUMBER(RRN)
RRDS
SKIP(number)
COUNT(number)
KSDS , ESDS, RRDS
SKIP refers to the number of input records to skip before beginning to copy
and COUNT specifies the number of output records to copy.
The REUSE Parameter With REPRO
The REUSE option of REPRO will logically delete the existing records of a
target KSDS, ESDS or RRDS and add new records from the source dataset as if the
target dataset were empty. In other words REUSE sets the value of the ending RBA
of the Highest Used control area in the data component (HURBA) to zero.
In order to use REUSE with REPRO it is a prerequisite that the target dataset
must have been defined with the REUSE option in the DEFINE CLUSTER command.
Merging DATASETS Using REPRO and REPLACE option
The REPRO command can be used for merging two datasets into one. The
target dataset can be a non-empty KSDS, ESDS or RRDS.
The input dataset can be any organization if the target dataset is a KSDS or
ESDS. However if the target dataset is an RRDS the source dataset has to be an
RRDS. KSDS to KSDS is the most common application of this merge technique.
If the REPLACE option is specified as part of the REPRO command then
records with duplicate primary keys (for a KSDS) and duplicate relative record
numbers (in the case of an RRDS) will be replaced.
IDCAMS - PRINT Command
It is used to print the contents of a dataset. The output can be made available
in various formats. This section deals with the PRINT command and its various
mutually exclusive options.
PRINT INDATASET(data set-name) | INFILE(dd-name)
options
PRINT - CHAR/ HEX/DUMP
This specifies the format in which to print.
CHARACTER(CHAR) prints as EBCDIC, 120 characters per line. This format is
mostly used for alphanumeric data. Any combination of bits that does not correspond
to a printable character will be printed as a period. If length of the record is greater
than 120 characters, it is printed in blocks of 120 characters per line.
HEX format prints each character in the dataset as two hexadecimal digits.
A maximum of 120 hexadecimal digits are printed on each line, an equivalent of 60
characters.
DUMP format is a combination of the character and hex formats. Each
character is printed both as a character and as the corresponding hex representation.
Each print line consists of 64 hex digits and the 32 related characters. If the record
Page:104
length is greater than 32 characters the rest of the record is printed in blocks of 64
hex digits and 32 corresponding characters per line.
PRINT SKIP, COUNT, FROM and TO
The records to be printed can be selected in the same way records are
selected in REPRO to COPY.
Where to start Printing
SKIP(number)
Where to stop Printing

COUNT(number)
FROMKEY(key-value)
FROMADDRESS(rba)
FROMNUMBER(rrn)
TOKEY(key-value)
TOADDRESS(rba)
TONUMBER(rrn)
Where Used
KSDS, ESDS, RRDS,
non-VSAM
KSDS, ALTERNATE INDEX
KSDS, ESDS
RRDS
A command that prints records 29, 30, and 31 in character format

PRINT INDATASET(MM01.CUSTOMER.MASTER)
CHARACTER
SKIP(28)
COUNT(3)
Empty file- check

If the file is empty, PRINT COUNT(1) ends with return-code as 4.
PRINT INDATASET(OUTPUT.DSN) CHARACTER COUNT(1)
IDCAMS - DELETE Command

The DELETE command can be used to delete both VSAM and non-VSAM
objects.
Syntax: DELETE ENTRY NAME OBJECT optional-parameters
OBJECT- CLUSTER GDG PATH AIX ALIAS CATALOG NONVSAM SPACE
USERCATALOG
Options ERASE/NOERASE PURGE/NOPURGE SCRATCH/NOSCRATCH
FILE FORCE/NOFORCE
DELETE- ENTRY NAME
The name of the entry to delete is coded. GENERIC NAMES can be also given.
DELETE - ERASE (ERAS) / NOERASE (NERAS)
When the DELETE command is executed it does not physically delete a
dataset. Only its ICF or VSAM catalog entry is removed making this space available
for future use. The data component remains on the disk until another dataset is
allocated on the same spot.
If the DELETE command is executed with the ERASE option, not only will the
entry be removed from the catalog but also the data component will immediately be
overwritten with binary zeroes. This option is coded for sensitive data files.
Page:105
DELETE - PURGE(PRG) / NOPURGE (NPRG)

NOPURGE specifies that the entry not to be deleted if the retention period has
not expired. (Default). So DELETE command dont delete the dataset before expiry
date. PURGE parameter must be coded to delete the dataset before retention period.
DELETE - SCRATCH(SCR) / NOSCRATCH(NSCR)
When a dataset is created, it will be listed in the catalog as well as the VTOC
of the DASD. SCRATCH specifies that the dataset to be removed from VTOC as well
as catalog. NOSCRATCH specifies that the dataset to be removed from the catalog
alone.
DELETE - FORCE(FRC)/NOFORCE(NFRC)
It specifies whether objects that are not empty should be deleted.
FORCE allows you to delete data spaces, generation data groups, and user
catalogs without first ensuring that these objects are empty.
NOFORCE causes the DELETE command to terminate when you request the
deletion of a data space, generation data group, or catalog that is not empty.
DELTE - FILE(DDNAME)
It specifies the name of the DD statement that identifies:
1. The volume that contains a unique data set to be deleted.
2. The partitioned data set from which a member (or members) is to be deleted.
3. The data set to be deleted if ERASE is specified.
4. The volume that contains the data space to be deleted.
5. The catalog recovery volume(s) when the entry being deleted is in a recoverable
catalog. If the volumes are of a different device type, concatenated DD statements
must be used. The catalog recovery volume is the volume whose recovery space
contains a copy of the entry being deleted.
IDCAMS - LISTCAT Command

The output of this command gives an insight into the inner functioning of
VSAM. LISTCAT is used to view dataset attributes, password & security information,
usage statistics, space allocation information, creation and expiration dates and
much more.
LISTCAT stands for LISTing a CATalog entry. It is useful for listing attributes
and characteristics of all VSAM and non-VSAM objects cataloged in a VSAM or ICF
catalog. Such objects can be the catalog itself, its aliases, the volumes it owns,
clusters, alternate indexes, paths, GDGs, non-VSAM files etc. The listing also
provides statistics about a VSAM object from the time of its allocation, namely the
number of CI and CA splits, the number of I/O on index and data components, the
number of records added, deleted and retrieved besides other useful information.
Syntax:
LISTCAT ENTRIES(OBJECT-NAME) ALL|ALLOCATION|VOLUME|HISTORY|NAME
Parameter
Meaning
NAME
HISTORY
VOLUME
ALLOCATION
ALL
Page:106
List the name and type of entry.

Lists reference information for the object including name, type of
entry, creation and expiration date and the release of VSAM under
which it was created.
Lists the device type and one or more volume serial number of the
storage volumes where the dataset resides. HISTORY information
is also listed.
Lists information that has been specified for space allocation
including the unit(cylinders, tracks etc.), number of allocated units
of primary and secondary space and actual extents. This is
displayed only for data and index component entries. If
ALLOCATION is specified VOLUME and HISTORY are included.
All the above details are listed
LEVEL Parameter in LISTCAT

LISTCAT ENTRIES(SMSXL86.*) ALL
This will return complete information about all objects having two levels of
qualification and having the first qualifier SMSXL86. But if we require information
about all objects having the high level qualifier SMSXL86, we have to use the LEVEL
parameter. LISTCAT LEVEL(SMSXL86.*) ALL
Example
//SMSXL861 JOB (36512),'MUTHU',NOTIFY=&SYSUID
// LISTCAT EXEC PGM=IDCAMS
//SYSIN
DD *
LISTCAT ENTRIES(SMSXL86.PAYROLL.MASTER) GDG
ALL
/*
IDCAMS - EXPORT and IMPORT commands

The EXPORT / IMPORT commands can be used for
1. Backup and recovery.
2. Exporting a dataset, an alternate index or a catalog to a different (but
compatible) system.
Advantages over REPRO
1. Catalog information is exported along with the data.
2. Cluster deletion and redefinition are not necessary during the import step
because input dataset already contains catalog information.
3. Also since the dataset contains catalog information it can be easily ported
to other systems. An exported dataset has cross-system portability.
But EXPORT / IMPORT can be used with VSAM datasets only and the dataset
created by EXPORT can have only a sequential organization. The dataset created by
the EXPORT step is not process-able until it has gone through a corresponding
IMPORT step. Thus performing an EXPORT/IMPORT takes up more time.
Page:107
Syntax: EXPORT entry-name OUTFILE(ddname) / OUTDATASET(dsname) optional parameters

Entry-name is the name of the object that need to be exported. OUTFILE mention
exported into what name.
EXPORT - INHIBITSOURCE|NOINHIBITSOURCE
It specifies whether the original data records (the data records of the source
cluster or alternate index) can be accessed for any operation other than retrieval
after a copy is exported. This specification can later be altered through the ALTER
command.
INHIBITSOURCE (INHS) - cannot be accessed for any operation other than retrieval.
NOINHIBITSOURCE - original data records in the original system can be accessed for
any kind of operation.
EXPORT - INHIBITTARGET(INHT)/NOINHIBITTARGET (NINHT)
It specifies how the data records copied into the target alternate index or
cluster can be accessed after they have been imported to another system. The ALTER
command is used to alter this parameter.
INHIBITTARGET - cannot be accessed for any operation other than retrieval.
NOINHIBITTARGET - Target object can be accessed for any type of operation after it
has been imported into another system.
EXPORT - TEMPORARY|PERMANENT
It specifies whether the cluster or alternate index to be exported is to
be deleted from the original system.
TEMPORARY specifies that the cluster or alternate index is not to be deleted
from the original system. The object in the original system is marked as temporary
to indicate that another copy exists and that the original copy can be replaced.
PERMANENT specifies that the cluster or alternate index is to be deleted from
the original system. Its storage space is freed. If its retention period has not yet
expired, you must also code PURGE. PERMANENT is the default.
EXPORT - ERASE|NOERASE
This specifies whether the data component of the cluster or alternate index to
be exported is to be erased or not (overwritten with binary zeros).
With ERASE specification, the data component is overwritten with binary zeros
when the cluster or alternate index is deleted.
With NOERASE specification, the data component is not overwritten with
binary zeros when the cluster or alternate index is deleted.
Example:
//EXPORT EXEC PGM=IDCAMS
//DD2
DD DSN=SMSXL86.LIB.KSDS.BACKUP(+!),
//
DISP=(NEW,CATLG,DELETE),UNIT=TAPE,
//
VOL=SER=121212,LABEL=(1,SL),
//
DCB=(RECFM=FB,LRECL=80)
Page:108
//SYSIN DD *
EXPORT A2000.LIB.KSDS.CLUSTER
OUTFILE (DD2)
/*
IDCAMS - IMPORT command

The IMPORT command restores the dataset from a backup copy created by
the EXPORT command. The INFILE parameter specifies the DDNAME that refers to
the backup copy created by the EXPORT command.
IMPORT INFILE(ddname)/ INDATASET(dsname) OUTFILE(ddname) / OUTDATASET(dsname)
optional parameters
IMPORT - INTOEMPTY (IEMPTY)

When this parameter is specified, a previously exported dataset can be
imported into an empty previously defined cluster. If this parameter is not specified
an attempt to import into an empty dataset will fail.
IMPORT - OBJECTS
When the OBJECTS parameter is coded the attributes of the new target
dataset can be changed. These attributes include VOLUMES and KEYRANGES.
IDCAMS-VERIFY Command
If a job terminates abnormally and a VSAM dataset is not closed, the catalog
entry for the dataset is flagged to indicate that the dataset may be corrupt(as the
end of file or last key is not updated in the index properly). In such case you would
get VSAM status code of 97 in the open of the file in the program. Before the
dataset can be opened again, the VERIFY command must be used to correctly
identify the end of the dataset and reset the catalog entry. Alternate solution is open
the file in File-aid in edit mode and just save it. It would update the index with end of
file.
Model syntax for the VERIFY command:
VERIFY {FILE(ddname[/password]) | DATASET(entryname[/password])}
Base cluster and alternate index can be verified. The verification of base cluster does
not verify its alternate indexes so each one of them must be treated separately.
IDCAMS - ALTER Command
The ALTER command is used to change many of the attributes of an existing
VSAM dataset. It is easier doing it this way than to delete the old dataset and to
redefine it with the new attributes. If the latter method is used the data will have to
be backed up before deletion to preserve the records in it and restored after
reallocation.
It is important to note that all the attributes cannot be changed with ALTER.
Few parameters that cant be altered are: CISZ, Type of Cluster,
IMBED/REPLICATE, REUSE/NOREUSE
Syntax: ALTER entry-name parameters
Page:109
Frequently ALTER is used for renaming (NEWNAME), altering free space

(FREESPACE), temporarily disabling updates over dataset (INHIBIT) and extend the
generation limit of GDG (LIMIT).
Example
JCL for an AMS job with comments that runs an ALTER command
//SMSXL861 JOB (36512),'Muthu',NOTIFY=&SYSUID
//NEWNAME EXEC PGM=IDCAMS
//SYSIN DD *
ALTER MM01.CUSTOMER.MASTER
NEWNAME(MM01.CUSTMAST) /* CHANGE NAME */ FREESPACE(10 10)
/* CHANGE FREESPACE ALLOCATION */
/*
SYSIN for Adding candidate Volumes
ALTER
EMPLOYEE.KSDS.DATA
ADDVOLUMES (VSAM03)
IDCAMS - MODAL COMMANDS

Functional commands are used to perform function. Modal commands are
used to test and SET the condition codes returned by the functional commands.
LASTCC contains the return-code of the last executed functional command
and if more than one functional command is executed in single IDCAMS step, then
MAXCC contains the maximum value of condition codes from the previously
executed functional commands. MAXCC is returned as step return code in JCL.
SET command is used to reset the MAXCC or LASTCC within AMS.
IF-THEN-ELSE statements are used to control the command execution
sequence.
In the below example, REPRO command loads data from a sequential dataset
on to a KSDS. Only if the condition code of the REPRO step is zero, the next LISTCAT
step will be executed. Otherwise the KSDS will be deleted. MAXCC is set to zero at
the end to avoid non-zero return code.
//SYSIN
DD *
REPRO INDATASET(SMSXL86.DATA.TEST) -
Page:110
OUTDATASET(SMSXL86.TEST.KSDS)
IF LASTCC = 0
THEN
LISTCAT
ENTRIES(SMSXL86.TEST.KSDS) ALL
ELSE
DELETE SMSXL86.TEST.KSDS
END-IF
SET MAXCC=0
/*
ICAMS step Return codes and meaning

Error
Code
0
4
8
12
16
Severity
of Error
No Error
Minor
Error
Major
Error
Logical
Error
Severe
Error
Meaning
Functional command completed its processing successfully
Processing is able to continue, but a minor error occurred, causing
a warming message to be issued
Processing is able to continue, but a more severe error occurred,
causing major command specifications to be bypassed
Generally, inconsistent parameters are specified, causing the
entire command to be bypassed
An error of such severity occurred that not only can the command
causing the error not be completed, the entire AMS command
stream is flushed
ALTERNATE INDEX
Need for AIX
1. KSDS dataset is currently accessed using a primary key. We want to access
the dataset RANDOMLY with another key.
2. We want to access ESDS file RANDOMLY based on key.
Steps Involved
Step1. Define AIX.
DEFINE ALTERNATEINDEX command defines an alternate index. Important
parameters of this command are:
Parameter
RELATE
Meaning
Relates AIX with base cluster
Page:111
NONUNIQUE/
UNIQUE
KEYS
UPGRADE
RECORDSIZE
Duplicates are allowed / not allowed in alternate key.

Defines the length and offset of alternate key in base cluster
Adds the AIX cluster to the upgrade set of base cluster. Whenever
base is modified, its upgrade set is also modified. UPGRADE is
default. NOUPGRADE didnt add the AIX to base cluster upgrade set.
Specifies the record size of alternate index record. It is calculated
using the formula in the next table.
AIX Record Size Calculation:

UNIQUE Alternate Key
KSDS
ESDS
5+Alternate Key length

+Primary Key Length
+RBA size
(RBA size = 4 bytes)
NONUNIQUE Alternate Key with

maximum n duplicates
+(n * Primary Key Length)
+(n * RBA size )
(RBA size = 4 bytes)
Five Byte control information of Alternate Index Record:

Byte-1
Byte-2
Byte-3
Byte-4
Byte-5
Type of Cluster; X00 indicates ESDS and X01 indicates KSDS.

Length of base cluster pointers in alternate index; Primary key
length for KSDS and X04 for ESDS.
Half word binary indicates number of occurrences of primary key pointers
in alternate index record. X0001 for unique alternate key.
Length of alternate key.
Step2. BLDINDEX
Alternate index should have all the alternate keys with their corresponding
primary key pointers. After the AIX is defined, this information should be loaded from
base cluster. Then only we can access the records using AIX. BLDINDEX do this LOAD
operation.
Important parameters of this command are:
1. INFILE and OUTFILE points to Base Cluster and Alternate index Cluster.
2. INTERNALSORT, EXTERNALSORT, WORKFILES:
Loaded AIX should have sorted alternate record keys with base cluster
keys. So an intermediate SORT is required in between reading the base cluster and
loading the AIX. We should allocate work datasets IDCUT1 and IDCUT2 for this
SORT.
Page:112
If INTERNALSORT parameter is coded, then AMS tried to do internal sort if

enough memory is available and it would be fast. If enough space is not available,
then if you have coded IDCUT1 and IDCUT2, it uses that space and does an external
sort.
If you code EXTERNALSORT parameter, then AMS wont try for internal sort.
If you want to give your own DDNAME files for sorting instead IDCUT1 and
IDCUT2, inform it to AMS by WORKFILE(DDNAME1, DDNAME2) parameter. Here,
DDNAME1 and DDNAME2 are allocated instead of IDCUT1 and IDCUT2.
Step3. Define Path
The DEFINE PATH command establishes a bridge between the base cluster and an
alternate index. It does not occupy any space. It is just a catalog entry, which
establishes a link between an alternate index and a base cluster. Important
parameters of this command are:
1. PATHENTRY command relates PATH with AIX.
2. When you open a PATH, respective base cluster will automatically be opened.
UPDATE opens the UPGRADE set of the base cluster also and it is default.
NOUPDATE opens only the base cluster and not its the UPGRADE set.
In BATCH, if you want to access base cluster with alternate key, you should
allocate the PATH in the JCL. The name of the DDNAME for the PATH in JCL is
DDNAME of the base cluster suffixed with 1 for first alternate key and n for nth
alternate key. Refer the COBOL material for more information.
In CICS, if you want access base cluster with alternate key, then you should
register the PATH as FCT entry.
Example:
1.Command that defines an alternate index
DEFINE AIX ( NAME(MMA2.EMPMAST.SSN.AIX)
RELATE(MMA2.EMPMAST)
KEYS(9 12)
UNIQUEKEY
UPGRADE
Page:113
REUSE
VOLUMES(MPS800) )
DATA
( NAME(MMA2.EMPMAST.SSN.AIX.DATA)
CYLINDERS(1 1) )
INDEX ( NAME(MMA2.EMPMAST.SSN.AIX.INDEX) )
2.Command that defines PATH

DEFINE PATH ( NAME(MMA2.EMPMAST.SSN.PATH)
PATHENTRY(MMA2.EMPMAST.SSN.AIX) UPDATE )
3.A job that builds an alternate index
//MM01LC3 JOB (36512),'MUTHU',NOTIFY=&SYSUID
//BLDINDX EXEC PGM=IDCAMS
//IDCUT1 DD UNIT=SYSDA,VOL=SER=MPS800,DISP=OLD
//IDCUT2 DD UNIT=SYSDA,VOL=SER=MPS800,DISP=OLD
//SYSIN
DD *
BLDINDEX INDATASET(MMA2.EMPMAST)
OUTDATASET(MMA2.EMPMAST.SSN.AIX) CATALOG(MMA2)
/*
The order of build index and definition of PATH does not matter.
AMP Parameter in JCL
The Access Method Parameter (AMP) completes information in an Access
method Control Block (ACB) for a VSAM data set.
The ACB can be coded for a key-sequenced (KSDS), entry-sequenced (ESDS),
or relative record (RRDS) VSAM data set. AMP is only supported for VSAM data sets.
AMP is most often used to allocate I/O buffers for the index and data components for
optimizing performance.
Syntax: AMP=(sub parameter, sub parameter,)
or
AMP=('sub parameter, sub parameter,')
Single apostrophes are required if the sub parameter(s) contain special characters.
AMORG
This parameter, which stands for Access Method ORGanization, indicates that
the particular DD statement refers to a VSAM dataset.
BUFND
This parameter gives the number of I/O buffers needed for the data
component of the cluster. The size of each buffer is the size of the data CI.
The default value is two data buffers one of which is used only during CI/CA splits.
Therefore the number of data buffers left for normal processing is one.
If more data buffers are allocated, then performance of sequential processing will
improve significantly.
Page:114
BUFNI
This parameter gives the number of I/O buffers needed for the index
component of the cluster. Each buffer is the size of the index. This sub-parameter
may be coded only for a KSDS because ESDS and RRDS do not have index
components. The default value is one index buffer.
If more index buffers are allocated, then performance of random processing
will improve significantly.
BUFSP
This parameter indicates the number of bytes for data and index component
buffers. If this value is more than the value given in the BUFFERSPACE parameter of
the DEFINE CLUSTER, it overrides the BUFFERSPACE. Otherwise BUFFERSPACE takes
precedence. The value of BUFSP is calculated as
BUFSP = DATA CISIZE x BUFND + INDEX CISIZE x BUFNI
However it is recommended not to code this parameter and let VSAM perform the
calculations from the BUFND and BUFNI values instead.
Other parameters are TRACE, STRNO, RECFM, OPTCD, CROPS.
If SMS is active, then VSAM datasets can be created in JCL without using IDCAMS as
below:
//KSDSFILE DD DSN=DEVI.CUST.MASTER,DISP=(NEW,CATLG,DELETE),
//
SPACE=(CYL,(10,10)),
//
LRECL=100,KEYOFF=10,KEYLEN=12,RECORG=KS
RECORG can also be ES(for Entry Sequenced Datasets), RR(for Relative Record
datasets) and LS(for Linear Datasets). Other parameters of DEFINE CLUSTER will be
assigned default values or you can additionally mention SMS parameter, DATACLASS
that is defined with predefined values.
DataBase 2
DB2
Page:115
Page:116
DB2 (Database 2)
History
IBMs first database is IMS. This is a hierarchical database. IBM introduces its second
base on relational concepts in 1980s and it is called as Database 2 (DB2).
Advantage of DBMS over File Management System
1.Increased Independency
If a new field is added to a file, the layout of the file should be changed in all
the programs with new field, though the programs do not use this field. But if a new
column is added to a table, it does not need any program change, as long as the
programs do not need this field. Thus programs using Databases are developed
independent of physical structure.
2.Less Redundancy
In file system, the same data is stored in multiple files to meet the
requirements of different applications. In DB2, we can store the common information
in one table and make all the applications to share it. By means of centralized control
and storing the fact at the right place and only once, data redundancy is reduced.
3.Increased Concurrency and Integrity
Concurrency allows more than one task to share a resource concurrently.
Integrity is defined as validity or correctness of data. In file system, if a file is
allocated with OLD disposition, then no other program can use it. DB2 allows more
than one task to use the same table in update mode. At the same time, integrity of
data is ensured by its sophisticated locking strategy.
4.Improved Security
In file system, all the records in the file are subjected to a specific access
right. The access cannot be controlled at field level or set level. But in DB2, we can
restrict access on column level and set level.
Types of Database
Hierarchical Relation between entities is established using parent-child
relationship-Inverted Tree Structure-This is the first logical model of DBMS. Data
stored in the form of SEGMENTS. Example: IMS
Network
Any entity can be associated with any other entity. So
distinguishing between parent and child is not possible. This is a linked structure
model. Data stored in RECORD and different record types are linked by SETS.
Example: IDMS
Relational
In mathematical terms, relation is TABLE. Data is stored in the form of

tables consists of Tuples (rows) and attributes (columns).
Example: DB2
Page:117
DB2 Objects
There are two types of elements or objects in DB2.

System Objects: objects that are controlled and used by DB2.
Example: DB2 Directory, DB2 catalog, Active & Archive logs, Boot Strap Dataset
(BSDS), Buffer Pools, Catalog visibility Database and Locks.
Data Objects: Objects that are created and used by the users.
Example: Tables, Indexes, Views, Table spaces, Databases, Storage Groups.
DB2 Object-Storage Group
Storage group is a collection of direct access volumes, all of the same type.
DB2 objects are allocated in Storage group. A storage group can have maximum 133
volumes.
Storage
Group 1
DASD001
DASD002
DASD003
DB2 Object-Table Space

Table Spaces are the physical space where the tables are stored. There are
three types of table spaces.
Simple Table Space:
More than one table is stored in the table space and single page can contain
rows from all the tables.
Table1 Row1
Table2 Row1
Table2 Row2
Page 1
Table1 Row2
Table1 Row3
Table3 Row1
Page 2
Table3 Row2
Table1 Row4
Table2 Row3
Page 3
Segmented Table Space:

Segment is logically contiguous set of n pages where n is defined by SEGSIZE
parameter of TABLESPACE definition. Tables are stored in one or more segments.
Mass delete and sequential access are faster in Segmented type table space.
Reorganization will restore the table in its clustered order. Lock table command locks
only the table and not the table space.
Table1 Row1
Table1 Row2
Table1 Row3
Segment 1
Table2 Row1
Table2 Row2
Table2 Row3
Segment 2
Table1 Row4
Table1 Row5
Table1 Row6
Segment 3
Page:118
Partitioned Table Space:

Only one table can be stored. 1-64 partitions are possible in table space.
NUMPART of TABLESPACE definition decides the partitions. It is partitioned with value
ranges for single or combination of columns and these columns cannot be updated.
Individual partitions can be independently recovered and reorganized. Different
partitions can be stored in different groups.
Table1 Row1
Table1 Row2
Table1 Row3
Partition 1
Table1 Row4
Table1 Row5
Table1 Row6
Partition 2
Table1 Row7
Table1 Row8
Table1 Row9
Partition 3
DB2 Object-Database
It is not a physical object. It is a logical grouping consists of Table spaces,

Index spaces, Views, Tables etc. Whenever you create an object you have to
explicitly say which Database it belongs to or DB2 implicitly assigns the object to the
right database.
In the creation of table, we should say in which database and in which table
Space we are housing the table (explicitly). In creating an index, we mention only
the table name. DB2 decides the database from the base table name.
It can also be defined as a unit of start and stop. There can be a maximum of
65,279 databases in a DB2 Subsystem.
DB2 Object - Table
Table is a structure consists of number of columns and rows. It is created

using the DDL CREATE TABLE.
Physical hierarchy:
ROWID is a pointer to data record in the page.
It is a 4byte field. (3 bytes for page and 1 byte for line in the page)
Page consists of 1 to 127 data records. It is the unit of I-O.
Table space consists of pages and it is stored in STOGROUP.
Database consists of one or more table space and other related objects.
DB2 Object- Index
Generally, Index is used to access a record quickly. There are two types of
Indexes: Clustering and Non-Clustering. If there is a Clustered index available for a
table, then the data rows are stored on the same page where the other records with
similar index keys are already stored on. There can be only one cluster index for a
table. Refer the diagram in the next page for better understanding of clustered and
non-clustered indexes.
Understanding: Root Page, Leaf Page and Non-Leaf page
Referring to VSAM basics, sequence set has direct pointers to control interval
and last level indexes have pointers to sequence set and higher-level indexes have
Page:119
pointers to lower level indexes and there is one root index from where the B-tree
starts.
Similarly Leaf-pages contain full keys with row-ids that points to data page. (Row id
has page number and line number). Non-leaf pages have pointers to leaf pages or
lower level non-leaf pages and the root page has pointer to first level non-leaf pages.
View of clustered index Diagram
*25 *50
*30 *42 *50
*7 *17 *25
*1 *3 *7
*11 *13 *17
Records for
keys 1 3 7
Records for
keys 11 13
17
Root Page
*19 *20 *25
Records for
keys 19 20
25
Non-Leaf Pages
Leaf Pages
Data Pages
View of Non-clustered index Diagram (non-leaf pages and data pages are shown)
*1 *3 *7
*11 *13 *17
Records for
keys 1 3
20
Records for
key 7 11
19
*19 *20 *25
Records for
keys 13 17
25
DB2 Object-Index Space
Unlike table spaces, Index spaces are automatically created when indexes are
created. Indexes are stored in index spaces.
DB2 Object - Alias & Synonym
Alias and Synonym are alternate name for a table. The main differences
between them are listed below:
SYNONYM
Private object. Only the user who
created it, can access it.
Used in local environment to hide
the high level qualifier.
When the base table is dropped,
associated synonyms are
automatically dropped.
SYSADM authority is not needed.
ALIAS
Global object. Accessible by anyone.
Used in distributed environment to hide the
location qualifier.
When base table is dropped, aliases are not
dropped.
To create an alias, we need SYSADM
Page:120
authority or CREATEALIAS privilege.
DB2 Object-View
Views provide an alternative way of looking at the data in one or more tables.
A view is a named specification of a result table. For retrieval, all views can be used
like base tables. Maximum 15 base tables can be used in a view.
Advantages of view
1.Data security: Views allows to set-up different security levels for the same base
Table. Sensitive columns can be secured from the unauthorized Ids.
2.It can be used to present additional information like derived columns.
3.It can be used to hide complex queries. Developer can create a view that results
from a complex join on multiple tables. But the user can simply query on this view
as if it is a separate base table, without knowing the complexity behind the building.
4.It can be used for domain support. Domain identifies a valid range of values that a
column can contain. This is achieved in VIEW using WITH CHECK OPTION.
Read Only Views
Views on which INSERT, UPDATE and DELETE operations cannot be carried

out, are called non-updateable views or read only views.
Characteristics of Updateable views
1.It should be derived from one base table and should not contain derived columns.
2.Views should not contain GROUP BY, HAVING and DISTINCT clauses at the
outermost level.
3.Views should not be defined over read-only views.
4.View should include all the NOT NULL columns of the base table.
DB2 Object- Catalog
It is a data dictionary for DB2, supporting and maintaining data about the
DB2 environment. It consists of 54 tables in the system database DSNDB06.
The data in DB2, about the data in the tables are updated when RUNSTATS utility
runs. The information in the DB2 catalog table can be accessed using SQL.
DB2 Object- Directory
This is second data dictionary of DB2 and used for storing detailed, technical
information about the aspects of DB2s operation. It is for DB2 internal use only.
It cannot be accessed using SQL statements. It is stored in the system database
DSNDB01.
DB2 Object- Active and Archive Logs
DB2 records all the data changes and significant events in active logs as and
when they occur. In case of failure, it uses this data to recover the lost information.
When active log is full, DB2 copies the content of active log to a DASD or magnetic
tape data set called archive log.
DB2 Object- Boot Strap Dataset
It stores the inventory of all-active and all-archive logs. During installation of
Page:121
DB2, two BSDS are created and kept in different volumes. BSDS datasets are VSAM
KSDS.
DB2 Object-Buffer Pools
Buffer Pools are virtual storage area, which DB2 uses for its temporary
purpose. There are fifty 4K buffer pools and ten 32K buffer pools in DB2.
DB2 Default buffers are BP0, BP1, BP2 and BP32.
Program Preparation
SQL (Structured Query Language) is the language that is used to

Retrieve / update / delete DB2 data. SQL statements are embedded into COBOL
Program within the scope of EXEC SQL and END-EXEC.
DB2 Program is first feed to DB2 Pre compiler that extracts the DB2
statements into DBRM and replace the source program DB2 statements with COBOL
CALL statements. This modified source is passed to COBOL compiler and then link
editor to generate load module. During pre compilation, time stamp token is placed
on modified source and DBRM.
On the other side, DBRM undergoes binding process and DB2 Optimizes
chooses the best access path for the extracted SQL statements and store it in PLAN.
Plan and Package
Binding process can happen in two stages, BIND PACKAGE and BIND PLAN.
One DBRM is created for one program. If the main program calls n number of
Sub-programs, then there will be n DBRMS in addition to main program DBRM.
These n + 1 DBRM can be directly feed to BIND PLAN to produce a single
PLAN or create m number of intermediate packages each formed by one or more
DBRM. This m numbers of packages are then feed to BIND PLAN step that produces
PLAN. Package is not executable but Plan is executable. To run a DB2 Program, Plan
is mandatory.
Advantages of Package:
1.When there is a change in sub program, it is enough to recompile
the subprogram and create the PACKAGE for that sub program. There is no need
for BIND PLAN.
2.Lck options and various bind options can be controlled at sub-program
level.
3.It avoids the cost of large bind.
4.It reduces Rebound time.
5.Versioning: Same package can be bounded with multiple programs with
different collection IDs.
Execution of DB2 Program
DB2 Program is executed using terminal monitor program IKJEFT01.

It can be executed using IKJEFT1A or IKJEFT1B also. They are alternate entry points
of IKJEFT01.DB2 region name, program to run, DB2 plan and PARM (if needed) are
provided in the SYSTSIN card of IKJEFT01.
//STEP EXEC PGM=IKJEFT01
//SYSPRNT DD SYSOUT=*
//SYSTSIN DD *
DSN SYSTEM(DST2)
== > In DB2 test region DST2
RUN PROGRAM(pgmname) PLAN(planname) == > Execute the program
Page:122
LIB(loadlibrary)
END
/*
== > whose load module is here!
Program Preparation in Detail

DCLGEN Library
(SYSLIB)
Modified Source (SYSCIN)*
IGYCRCTL
COBOL COMPILER
Object Module
(SYSLIN)
IEWL
LINK EDITOR
Load Module *
(SYSLMOD)
COBOL-DB2 PROGRAM
(SYSIN)
DSNHPC
DB2 pre-compiler
COPYBOOK Library
(SYSLIB)
COMPILER Options
(PARM)
SUBROUTINE Library
(SYSLIB)
Link Edit Options
(PARM)
Pre-compiler Options
(PARM)
Database Request
Modules (DBRM)*
IKJEFT01
(BIND PACKAGE)
BIND
PARAMETERS
(SYSTSIN)
IKJEFT01
LAN
(BIND
PLAN)
PLAN *
* Time Stamp token is passed from pre-compilation stage to Load module and Plan.
When first time the plan is accessed, the time stamp of plan is verified against
the time stamp of load module. If they match, then table will be accessed using the
access path in PLAN and proceed further. If they dont match, then the program
abnormally ends with SQLCODE of 818. It means the load module and plan are out
Page:123
of sync. It usually occurs if we modify the program and pre-compile, compile and link
edit but did not bind the plan with latest DBRM produced in pre-compilation stage.
There is a modification in COBOL part. There is no change in any of the DB2
statements in the program. Is binding necessary?
Yes. It is necessary. For COBOL changes, we have to compile and link edit.
For compilation we need pre-compiled source. So new time stamp is transferred to
your load module from the pre-compilation stage. If we dont BIND it, timestamp
mismatch is obvious and the program will abnormally end.
Can this time stamp check be avoided?
If the program pre-compiled with LEVEL option, then there wont be time stamp
verification. But this option is not generally recommended.
Bind Card
BIND PACKAGE sub command is used to bind a DBRM to a package and BIND
PLAN is used to bind a DBRM to a plan or group of packages to a PLAN.
Sample Bind Card:
//BIND EXEC PGM=IKJEFT01
//SYSTSIN DD *
DSN
BIND PLAN(EMPPLAN)
MEMBER(EMPDBRM)
VALIDATE(BIND)
ISOLATION(CS)
RELEASE(COMMIT)
OWNER(SMSXL86)
LIB(SMSXL86.DB2.DBRMLIB)
/*
Bind Parameter MEMBER and LIBRARY
DBRM to be bound is given in MEMBER and the partitioned dataset containing
the DBRM is given in LIB. During BIND Package, if an existing package is used to
create the new package then, COPY should be mentioned instead of MEMBER and
LIBRARY.
COPY(collection-id.package-id)
COPYVER(version-id) version of the copy to be used. Default is empty if omitted.
Bind Parameter PKLIST
PKLIST is a BIND parameter of BIND PLAN. Packages to be connected with
PLAN are named here. PKLIST (COLL1.*) bound all the packages with collection ID
COLL1 to the PLAN.
Package Naming Collection-ID and Versioning
Packages have three qualifiers. Location ID is used in distributed
environment. Collection ID is used for grouping of packages.
Naming Syntax: Location-id. Collection-id. Package-name
Page:124
VERSION. With packages, the pre-compiler option VERSION can be used to bind
multiple versions of the programs into package.
PROG-------- > Pre-compile ----------- BIND Package --------COLL1.PROG.TEST
With VERSION(TEST)
-------- > Pre-compile ----------- BIND Package --------COLL1.PROG.PROD
With VERSION(PROD)
Bind Parameter ACTION
Package or Plan can be an addition or replacement. Default is replacement.
REPLACE will replace an existing package with the same name, location and
collection. REPLVER(version-id) will replace a specific version of an existing package.
ADD Add the new package to SYSIBM.SYSPACKAGE.
Syntax: ACTION(ADD|REPLACE)
Bind Parameter Isolation Level
The isolation level parameter of the BIND command describes to what extent a
program bound to this package can be isolated from the effects of other programs
running. This determines the duration of the page lock.
CURSOR STABILITY (CS) As the cursor moves from the record in one page to the
record in next page, the lock over the first page is released (provided the record is
not updated). It avoids concurrent updates or deletion of row that is currently
processing. It provides WRITE integrity.
REPEATABLE READ (RR) All the locks acquired are retained until commit point.
Prefer this option when your application has to retrieve the same rows several times
and cannot tolerate different data each time. It provides READ and WRITE integrity.
UNCOMMITTED READ (UR) It is also known as DIRTY READ. It can be applied only
for retrieval SQL. There are no locks during READ and so it may read the data that is
not committed. Highly dangerous and use it when concurrency is your only
requirement. It finds its great use in statistical calculation of the large table and
data-warehousing environment.
Bind Parameter ACQUIRE and RELEASE
ACQUIRE(USE) and RELEASE(COMMIT) DB2 imposes table or table space
lock when it executes an SQL statement that references a table in the table space
and it release the acquired lock on COMMIT or ROLLBACK. This is the default option
and provides greater concurrency.
ACQUIRE(USE) and RELEASE(DEALLOCATE) Locks table and table spaces on
use and releases when the plan terminates.
ACQUIRE(ALLOCATE) and RELEASE(DEALLOCATE) When DB2 allocates the
program thread, it imposes lock over all the tables and table spaces used by the
program. This option avoids deadlocks by locking your source at the beginning but it
reduces concurrency.
ACQUIRE(ALLOCATE) and RELEASE(COMMIT) is NOT ALLOWED. This
increases the frequency of deadlocks.
Bind Parameter SQLERROR
Page:125
It says whether to create the package in case of SQL errors.

SQLERROR (NOPACKAGE) Default - If there is any SQL error, package will not be
created.
SQLERROR (CONTINUE) Create the Package even if SQL errors are encountered.
Bind Parameter VALIDATE

It controls the handling of object not found and not authorized errors.
Default is VALIDATE (RUN) If all objects are found and all privileges are held, then
dont check once again at execution time. If there is any privilege issue or object
not found error, then produce WARNING messages, bind the package and check
authorization and existence at execution time.
VALIDATE (BIND) If there is any privilege or object-not-found issue, then produce
ERROR message and create the package only if SQLERROR(CONTINUE) is coded in
bind card. The statement in error will not be executable.
BIND Parameter EXPLAIN
EXPLAIN(YES) loads the access path selected by the optimizer in
PLAN_TABLE. EXPLAIN(NO) is default and it wont load any such details in
PLAN_TABLE.
BIND Parameter QUALIFIER
Qualifiers are added at the bind time to the objects used in the program.
It is suggested NOT to use any qualifiers in the program and use this parameter
during BIND so that the program can be easily promoted from one region to next
region without any change in program.
REBINDING
When SQL statements are not changed but a new index is added or
RUNSTATS is run, then it is advisable to do a REBIND for the best possible access
path. REBIND PLAN is cheaper than BIND with ACTION (REPLACE).
SYSTSIN CARD should be
REBIND PLAN(PLAN-NAME)
VALIDATE(BIND)
..
Page:126
Host Variables
DB2 is separate subsystem. As DB2 data are stored in external address space,
you cannot directly modify or read from your program. During read, the DB2 values
are retrieved into your working storage variables. During update, the DB2 columns
are updated using your working storage variables. These working storage variables
used for retrieving or updating DB2 data should be of same size of DB2 columns.
They are called as host variables as they are defined in the host language (COBOL).
Host variables are prefixed with colon ( : ) in the embedded SQL.
DCLGEN (Declaration Generator)
DCLGEN is a DB2 utility. If we provide table details, it will generate DB2
Structure and host variable structure of the table. It can also generate NULL
indicators for the NULL columns. This output is stored in a PDS member. These
members are included in the program using the INCLUDE <pds-member>. INCLUDE
is a pre-compiler statement and so it should be coded within the scope of EXEC SQL
and END-EXEC.
DCL generated host variable names are same as DB2 column names.
As underscore is not a valid character for variable names in COBOL, it is replaced by
hyphen in host variable generation of DCLGEN. Prefix and suffix can be added to all
the variables while creating the DCLGEN copybook.
Host variables can be declared in working storage section. But they cannot be
stored in copybooks as other file layouts.
INCLUDE is expanded during pre-compilation time but COPY is expanded
during compilation and so declaration of host variable in copybook will produce errors
during pre-compilation.
DECLARE TABLE, that is table structure of DCLGEN is NOT really needed for
pre-compilation. But if it is used, then any misspelled table name, column names are
trapped in pre-compilation stage itself.
DB2 need not be up during pre-compilation. As pre-compiler just extracts the
DB2 statements and produce DBRM, it does not need DB2 connection. But DB2
region should be up for binding as the optimizer builds access path in this stage from
catalog information.
Page:127
DCLGEN Panel (Option 2 of DB2I)

_______________________________________________________________________________
|
|
| DSNEDP01
DCLGEN
SSID: DSN
|
| ===>
|
|
|
| Enter table name for which declarations are required:
|
|
1 SOURCE TABLE NAME ===> DSN8410.EMPTABLE
(Unqualified table name)
|
|
2 TABLE OWNER
===>
(Optional)
|
|
3 AT LOCATION ..... ===>
(Optional)
|
|
|
| Enter destination data set:
(Can be sequential or partitioned)
|
|
4 DATA SET NAME ... ===> SMSXL86.DB2.INCLUDE(EMPTABLE)
|
|
5 DATA SET PASSWORD ===>
(If password protected)
|
|
|
| Enter options as desired:
|
|
6 ACTION .......... ===> ADD
(ADD new or REPLACE old declaration)
|
|
7 COLUMN LABEL .... ===> NO
(Enter YES for column label)
|
|
8 STRUCTURE NAME .. ===>
(Optional)
|
|
9 FIELD NAME PREFIX ===>
(Optional)
|
| 10 DELIMIT DBCS .... ===> YES
(Enter YES to delimit DBCS identifiers)
|
| 11 COLUMN SUFFIX ... ===> NO
(Enter YES to append column name)
|
| 12 INDICATOR VARS .. ===> NO
(Enter YES for indicator variables)
|
|
|
|
|
| PRESS: ENTER to process
END to exit
HELP for more information
|
|
|
|__________________________________________________________________________________|
DSNE905I EXECUTION COMPLETE, MEMBER EMPTABLE ADDED

DB2 Data Types and equivalent host variables.
DB2 Column
SMALLINT
INTEGER
DECIMAL (p, q)
(P should be less than 32)
Bytes
2
4
Int
(P/2)
COBOL PIC Clause

PIC S9(04) COMP
PIC S9(09) COMP
PIC S9(p-q)V9(q)
DATE
PIC X(10)
Bytes
2
4
Integer
((P+Q)/2
+1)
8
Page:128
TIME
TIMESTAMP
3
10
PIC X(08)
PIC X(26)
6
26
CHAR(n) (n=1 to 254)

VARCHAR(n) (n=0-4046)
N
N
PIC X(n)
01 WS-COLUMN.
N
N+2
yyyy-mm-dd-hh.mm.ss.nnnnnn
49 WS-COLUMN-LENGTH PIC S9(04)

COMP
49 WS-COLUMN TEXT PIC X(n)
Limit of Date: Jan 1st 1 ad Dec 31st 9999 AD

Limit of Time: 000000 240000
Limit of Time Stamp: 00010101000000000000 00001231240000000000
SQLCA
SQLCA is SQL Communication area. It is a DB2 copybook that should be

included in all the DB2 programs. The fields declared in the copybook are updated for
every DB2 query. The Length of SQLCA is 136. The most important field in this
copybook is SQLCODE. The success or failure of DB2 query can be checked in this
field.
Value of 0 in SQLCODE indicates the successful operation.
Positive value in SQLCODE indicates the completed operation with some exception
Negative value in SQLCODE indicates the unsuccessful operation.
So it is common programming practice that the SQLCODE of every query
must be verified for valid values after the execution of the query. If the SQLCODE
value is not acceptable, then the control is transferred to ABEND routine that would
end the program with proper error messages.
SQLERRD (3) contains the number of rows affected by a DB2 INSERT/DELETE
operation. If the operation is DELETE, this contains number of rows deleted after the
execution of the query.
WHENEVER statement
WHENEVER is a error-trapping statement that directs the processing to
continue or branch to an error handling routine based on the SQLCODE returned for
the statement. When it processed, it applies to all the subsequent SQL statements
issued by the application program.
EXEC SQL
WHENEVER condition action
END-EXEC.
Conditions can be NOT FOUND, SQLWARNING or SQLERROR.
Action can be either CONTINUE or GO TO.
NOT FOUND is TRUE if SQLCODE is 100.
SQLWARNING is TRUE if SQLCODE is greater than zero but not equal to 100
Page:129
SQLERROR is TRUE if SQLCODE is less than zero.

It is almost always safe to code specific SQLCODE checks after each SQL
statement and process accordingly. Avoid using WHENEVER.
DSNTIAR SQLCA Formatter
DSNTIAR is an IBM supplied program, which would display the error
messages in a formatted way. Most of the installations call this program in their error
routine.
CALL DSNTIAR USING SQLCA, ERROR-MESSAGE, ERROR-TEXT-LEN.
01 ERROR-MESSAGE.
05 ERROR-LEN PIC S9(04) COMP VALUE +1320.
05 ERROR-TEXT PIC X(132) OCCURS 10 TIMES.
77 ERROR-TEXT-LEN PIC S9(09) COMP VALUE +72.
Components of DB2
DB2 internal structure is very complex and contains large number of
components. But from a high-level point of view, DB2 can be considered to have four
major components.
System Services component: Supports system operation, operator communication,
logging and similar function.
Locking services component: Provides the necessary controls for managing
concurrent access to data.
Database services component: Supports the definition, retrieval and update of user
and system data.
Distributed data facility component: Provides DB2s distributed database support.
Before get into SQL, let us briefly see the sub- components of Database services
component.
Database services component
It has five sub-components in it. We have already discussed the functions of
two components Pre-compiler and Optimizer (Bind)
Runtime supervisor:
It is resident in the main memory when the application program is executing.
Its job is to oversee that execution. When the program requests some database
operation to be performed, control first goes to the runtime supervisor, which uses
the control information in the application plan to request the appropriate on the part
of the Data Manager.
Data Manager (DM):
Page:130
It performs all the normal access method functions search, retrieval, update
etc. It invokes other system components in order to perform detail functions such as
locking, logging etc.
Buffer Manager (BM):
It is responsible for transferring the data between external storage and virtual
memory. It uses sophisticated techniques such as read-ahead-buffering and lockaside-buffering to get the best performance out of the buffer pools under its control
to minimize the amount of physical i/o actually performed.
Catalog and Directory are also part of Database services component only.
SQL
SQL is fourth generation language. The definition of fourth language is Tell
the need. System will get it done for you . There is no need to instruct the method
to get it.
For example, I want list of employees working currently in Malaysia. If the
data is in file, then I have to declare, define, and Open the file. Then I should read all
the records into my working storage one by one and if their location is Malaysia, I
display them. Finally I close the file.
If the data is in DB2, I would simply write a query (SELECT * FROM table
where location=Malaysia) and DB2 optimizer do everything for me in identifying the
best access path. No worries of DECLARE, DEFINE, OPEN, READ and OMIT stuffs.
Thats the power of fourth generation language and that makes the programmer life
easy!!
SQL has three kinds of statements.
DDLData Definition Language Statements
CREATE ALTER DROP
DMLData Manipulation Language Statements SELECT INSERT UPDATE DELETE
DCLData Control Language Statements
GRANT REVOKE
DDL-CREATE
This statement is used to create DB2 objects.
General Syntax
CREATE object object-name parameters
Table creation Sample
CREATE TABLE table-name
(Column definitions,
Primary Key definition,
Foreign Key definition,
Alternate Key definition,
LIKE table-name /View-name,
Page:131
IN DATABASE-NAME.TABLESPACE-NAME)
DDL-Column Definition
Define all the columns using the syntax:
Column-name Data-type Length (NULL/NOT NULL/ NOT NULL WITH DEFAULT)
Data-types are already explained in DCLGEN section.
Column-name can be of maximum length 30. It should start with an alphabet and
it can contain numbers and underscore.
NULL means UNKNOWN, if the value is not supplied during an insertion of a row,
then NULL will be inserted into this column.
NOT NULL means, value for this column should be mentioned when inserting a
row in the table. NOT NULL with DEFAULT means, if the value for this column is
not supplied during an insertion, then based on type of the column default values
will be moved into the table. Default values are listed in the below table.
Data- Type
CHAR
VARCHAR
DATE, TIME, TIMESTAMP
INTEGER SMALLINT DECIMAL
Default Value
Spaces
Empty String (String of length 0)
Current (DATE/TIME/TIMESTAMP)
Zero
DDL-Primary Key definition

Candidate key is defined as column(s) that uniquely identifies each row in the
table. One of the candidate keys is defined as primary during the table creation. The
main entity, about which the information is stored, is selected as primary key.
If the PRIMARY KEY is single column, then it can be coded along with column
definition. If the primary key is composite key (more than one column), then it can
be defined after the column definition. Primary key definition can be added or
dropped later using ALTER table.
The definition of primary key is complete only after unique index is created on
it. Until that time, the table is unavailable for use. If the primary key is added using
ALTER command, then unique index on the key column(s) must already be available.
Syntax: PRIMARY KEY (column1, column2)
DDL-Foreign Key definition
Defining a foreign key establishes a referential constraint between the foreign
key (Child Table) and the primary key (Parent table). The parent table of a referential
constraint must have a primary key and a primary index.
There must be a value in parent table (primary key) for every non-null value
in child table (foreign key). Foreign key can be defined along with column definition
or separately after the definition of all the columns.
Syntax 1: (Along with column definition)
COLUMN1 CHAR(4) REFERENCES TABLE2 ON DELETE CASCADE/DELETE/SET NULL
Syntax 2: (After column definition)
FOREIGN KEY FKEY(COLUMN1,COLUMN2) REFERENCES TABLE2 ON
DELETE/CASCADE/SET NULL
Page:132
Delete Rule
Delete rule defines what action needs to be taken on child table when there is
a deletion in primary table. The three possibilities are below:
CASCADE: Child table rows referencing parent table will also be deleted.
RESTRICT: Cannot delete any parent row when there is reference in child table.
SET NULL: SET the foreign key value as NULL when the respective primary row is
deleted.
Insert and Update rules cannot be defined. But they can be implemented using
TRIGGERS.
DDL-Alternate Key Definition
Candidate keys that are not defined as primary key are called alternate keys.
This can be defined along with column definition or separately after the definition of
all the columns using the keyword UNIQUE. Alternate key definitions CANNOT be
added later using ALTER command.
Syntax 1: COLUMN1 CHAR(10) UNIQUE NOT NULL;
Syntax 2: UNIQUE(Column1)
DDL- Adding Constraints
A check constraint allows you to specify the valid value range of a column in
the table. For example, you can use a check constraint to make sure that no salary in
the table is below Rs10000, instead of writing a routine in your application to
constrain the data.
CHECK clause and the optional clause CONSTRAINT (for named check
constraints) are used to define a check constraint on one or more columns of a table.
A check constraint can have a single predicate, or multiple predicates joined by AND
or OR. The first operand of each predicate must be a column name, the second
operand can be a column name or a constant, and the two operands must have
compatible data types.
Syntax 1: ADD CHECK (WORKDEPT BETWEEN 1 and 100);
Syntax 2: ADD CONSTRAINT BONUSCHK CHECK (BONUS <= SALARY);
Although CHECK IS NOT NULL is functionally equivalent to NOT NULL, it
wastes space and is not useful as the only content of a check constraint. However,
later if you want to remove the restriction that the data be Non-null, you must define
the restriction using the CHECK IS NOT NULL Clause.
DDL-Index creation
Table spaces should be created before creating the table. But index spaces
are automatically created with creation of index. So create index command provides
optional parameters of USING STOGROUP, PRIQTY, SECQTY, FREEPAGE and PCTFREE
etc. for the allocation of index space. But usually we dont mention these parameters
and allow the system to allocate it in the right place with respect to table space.
CREATE [UNIQUE] INDEX index-name on Table-name (column-name [asc/desc],..)
Page:133
CREATE UNIQUE INDEX EMPINDEX ON EMPTABLE (EMPNO ASC)

Creating unique index on a NULL column is possible if there is only one NULL
is available in that column.
DDL-Sample Storage group Creation
CREATE STOGROUP DSN8G41D
VOLUMES (DBPK01, DBPK02)
VCAT DB2CATZ;
Using the ALTER command, volumes can be added or removed at later stage
using the command ADD VOLUMES and REMOVE VOLUMES.
DDL-Sample Database Creation
CREATE DATABASE DSN8D41A
STOGROUP DSN8G410
BUFFERPOOL BP0;
DDL-Sample Table space Creation

CREATE TABLESPACE DSN8S41D
IN DSN8D41A
USING STOGROUP DSN8G410
PRIQTY 20
SECQTY 20
ERASE NO
LOCKSIZE PAGE LOCKMAX SYSTEM
BUFFERPOOL BP0
CLOSE NO
FREEPAGE 0 PCTFREE 5
DSETPASS DSN8;
PRIQTY Primary space to be allocated during table space creation. (KB)
SECQTY -Secondary space taken as amount of data in table space grows. (KB)
ERASE Whether DB2 defined datasets are to be erased when table space is
dropped.
LOCKSIZE Indicates size of lock Can be TABLESPACE, TABLE, PAGE, ROW or ANY.
ANY means DB2 will decide the lock.
LOCKMAX Indicates maximum number of page / row locks above which lock
escalation occurs.
BUFFERPOOL Buffer pool associated with the table space.
CLOSE Indicates whether datasets associated with table space should be closed
when there are no current users of table space.
FREEPAGE Number of pages after which a empty page is available.
PCTFREE - Percentage of page to be left free for future inserts.
Page:134
DDL-Sample Table Creation

CREATE TABLE DSN8410.EMP
(EMPNO
CHAR(6)
NOT NULL,
NAME
VARCHAR(50) NOT NULL,
WORKDEPT CHAR(3),
PHONENO
CHAR(4)
CONSTRAINT NUMBER CHECK
(PHONENO >= 0000 AND
PHONENO <= 9999),
BIRTHDATE DATE,
SALARY
DECIMAL(9,2),
PRIMARY KEY (EMPNO),
FOREIGN KEY RED (WORKDEPT) REFERENCES DSN8410.DEPT
ON DELETE SET NULL)
IN DSN8D41A.DSN8S41D;
DDL-Sample View Creation

The general syntax of view definition is
CREATE VIEW view-name (column1,..) AS sub-query [WITH CHECK OPTION]
WITH CHECK OPTION ensures that all the inserts and updates made to the
base table using this view satisfies the limit imposed by the where condition of the
sub-query. (Domain Integrity)
CREATE VIEW DB2GROUP (EMPID EMPNAME SKILLSET)
AS SELECT EMPID, EMPNAME, SKILLSET
FROM DSN8410.EMP
WHERE SKILLSET = DB2 WITH CHECK OPTION;
This view shows name and id of all the employees who knows DB2.
WITH CHECK OPTION ensures that all the inserts made to base table using the view
DB2GROUP, must have DB2 as SKILLSET. Any update-attempt to change the value of
SKILLSET from DB2 to anything else using the view DB2GROUP is restricted.
DDL-ALTER TABLE
When you create a table using model table, the columns are defined in the
new table as in the model table. But the PRIMARY, FOREIGN KEY definitions are not
inherited. We have to define them using ALTER TABLE. Before defining primary key
using ALTER TABLE, there should be unique index.
CREATE TABLE YDEPT
Page:135
LIKE DSN8410.DEPT;
CREATE UNIQUE INDEX YDEPTX
ON YDEPT (DEPTNO);
ALTER TABLE YDEPT
PRIMARY KEY(DEPTNO);
Using ALTER statement, alternate keys cannot be defined. Other than that we
can add columns, constraints, checks & DROP primary key.
When a new column is added with NOT NULL WITH DEFAULT, then the value for
the column for all the rows already exist in the table is filled based on the type of
column:
1. Numeric
Zero
2. Char, Varchar
-Spaces
3. Date
-01/01/0001
4. Time
-00:00 AM
Using ALTER command, we can extend the length of VARCHAR/CHAR items, switch
the data-type of a column within character data-types (CHAR/VARCHAR);numeric
data types(SMALLINT,INTEGET,REAL,FLOAT,DOUBLE,DECIMAL) and within graphical
datatypes(GRAPHIC,VARGRAPHIC).
ALTER TABLE DSN8410.EMP
ALTER COLUMN WORKDEPT
SET DATATYPE CHAR(6);
DDL-DROP
DROP deletes the table definition in addition to all the rows in it. You also lose
all synonyms, views, indexes, referential and check constraints associated with that
table. Plans and Packages are deleted using the command FREE.
DROP Object Object-name.
DROP TABLE table-name
FREE PACKAGE(COLL1.*) - Deletes all the packages belongs to collection COLL1.
Data Manipulation Language statements
DML-SELECT
SELECT statement is the heart of all queries and it is specifically used for
retrieving information. As a primary statement to retrieve the data, it is the most
complex and most complicated DML statement used. Six different clauses can be
used with a SELECT statement and each of these clauses has its own set of
predicates. They are FROM-WHERE-GROUPBY-HAVING-UNION and ORDER BY.
Syntax: SELECT columns FROM tables
WHERE conditions
GROUP BY columns
HAVING conditions
ORDERBY columns
Up to 750 columns can be selected. 15 sub-queries can be coded in addition to one
main query.
Page:136
DML-FROM clause of SELECT

FROM clause is used, in conjunction with one or more references, to specify
DB2 data manager where to retrieve data.
WHERE Clause
It is always followed by a search condition that contains one or more
predicates that define how the DB2 DM is to choose the information that will be
contained in the result dataset produced.
Any relational operator can be coded. (<>, >. < , = ,<=, >=)
DML-SELECT-WHERE- NOT Condition
NOT keyword can be used to select all the rows except the row identified with
the search condition. Syntax: WHERE NOT <condition>.
NOT cannot be used directly with comparison operators. NOT = is invalid.
DML-SELECT-WHERE-LIKE condition
LIKE is used for pattern search. % is a wild card for any number of zero or
more characters and _ is used to indicate single unknown character.
If you want to look for % in the string, then you should define ESCAPE
character. Ex: WHERE col1 LIKE MUTHU_+%SARA% ESCAPE +
This will look for a string that starts with MUTHU_%SARA where _ can be
any character.
DML-SELECT-WHERE-BETWEEN and IN
Use BETWEEN to check the value of column is within the range. The lower and
upper boundaries are INCLUSIVE. Use IN to check the value of column against a
Values list.
DML-SELECT-WHERE-ANY/SOME/ALL
SELECTWHERE COLUMN1 > ANY|ALL|SOME (Sub-Query)
ALL - If the value of COLUMN1 is greater than all the values return by subquery, then only the outer table row will be selected. If no rows are returned by subquery then all the rows of outer table will be selected.
ANY and SOME If COLUMN1 value is greater than one of the values return
by sub-query, then the outer table row will be selected.
DML-SELECT-WHERE-EXIST
This is useful when you want to just check whether one or more rows exist.
Syntax: WHERE EXISTS (sub-query)
Your main query will return result only if at least one row exists for your sub
query. NOT EXISTS check for unavailability of any rows for the search condition in
the sub query.
DML-SELECT-GROUP BY
Use GROUP BY for grouping rows by values of one or more columns. You can
then apply column function for each group. Except for the columns named in the
GROUP BY clause, the SELECT statement must specify any other selected columns as
an operand of one of the column functions.
Page:137
If a column you specify in the GROUP BY clause contains null values, DB2
considers those null values to be equal. Thus, all nulls form a single group.
You can also group the rows by the values of more than one column
DML-SELECT-HAVING (Subjecting Group to conditions)
Use HAVING to specify a search condition that each retrieved group must
satisfy. The HAVING clause acts like a WHERE clause for groups, and contain the
same kind of search conditions you specify in a WHERE clause.
The search condition in the HAVING clause tests properties of each group
rather than properties of individual rows in the group.
DML-SELECT-UNION and UNION ALL
Using the UNION keyword, you can combine two or more SELECT statements
to form a single result table. When DB2 encounters the UNION keyword, it processes
each SELECT statement to form an interim result table, and then combines the
interim result table of each statement. If you use UNION to combine two columns
with the same name, the result table inherits that name.
You can use UNION to eliminate duplicates when merging lists of values
obtained from several tables whereas UNION ALL retains duplicates.
DML-SELECT-ORDER BY
The ORDER BY clause is used to sort and order the rows of data in a result
dataset by the values contained in the column(s) specified. Default is ascending
order (ASC). If the keyword DESC follows the column name, then descending order
is used. Integer can also be used in place of column name in the ORDER BY clause.
SELECT COL1,COL2,COL3 FROM TAB1 ORDER BY COL1 ASC COL3 DESC.
SELECT COL1,COL2,COL3 FROM TAB1 ORDER BY 1 ASC 3 DESC.
DML-SELECT-CONCAT
CONCAT is used for concatenating two columns with the string you want in between.
To concatenate the last name and first name with comma in between,
SELECT LASTNAME CONCAT ',' CONCAT FIRSTNME
FROM DSN8410.EMP;
COLUMN FUNCTION AND SCALAR FUNCTION
Column function receives a set of values from group of rows and produces a single
value. SCALAR function receives one value and produces one value.
COLUMN FUNCTION:
Function
MAX
MIN
AVG
SUM
COUNT
Returns-Maximum value in a column

Minimum value in a column
Average value of a column
Sum value of a column
Number of selected rows
Example
MAX(SALARY)
MIN(SALARY)
AVG(SALARY)
SUM(SALARY)
COUNT(*)
Page:138
1. DISTINCT can be used with the SUM, AVG, and COUNT functions. The selected
function operates on only the unique values in a column.
2. SUM and AVG cannot be used for non-numeric data types whereas MIN, MAX and
COUNT can.
SCALAR FUNCTION:
Function
CHAR
DATE
DAY
DAYS
MONTH
YEAR
DECIMAL
FLOAT
HEX
INTEGER
HOUR
MINUTE
SECOND
MICROSECOND
TIME
TIMSTAMP
LENGTH
SUBSTR
VALUE
Returns--String representation of its first

argument
Date derived from its argument
Day part of its argument
Integer representation of its
argument
Month part of its argument
Year part of its argument
Decimal representation of its
first argument
Floating-point representation of
its first argument
Hexadecimal representation of
its first argument
Integer representation of its
argument
Hour part of its argument
Minute part of its argument
Second part of its argument
Microsecond part of its
argument
Time derived from its argument
Timestamp derived from its
argument or arguments
Length of its argument
Sub-string of a string
The first argument that is not
null
Example
CHAR(HIREDATE)
DATE(1977-12-16)
DAY(DATE1-DATE2)
DAYS(1977-12-16)
-DAYS(1979-11-02) + 1
MONTH(BIRTHDATE) = 16
YEAR(BIRTHDATE)=1977
DECIMAL(AVG(SALARY),8,2)
FLOAT(SALARY)/COMM
HEX(BCHARCOL)
INTEGER(AVG(SALARY)+0.5)
HOUR(TIMECOL) > 12
MINUTE(TIMECOL) > 59
SECOND(TIMECOL) > 59
MICROSECOND(TIMECOL) > 12
TIME(TSTMPCOL) < 12:00:00
TIMESTAMP(DATECOL,TIMECOL)
LENGTH(ADDRESS)
SIBSTR(FSTNAME,1,3)
VALUE(SMALLINT1,100) +
SMALLINT2 > 1000
CURSOR
CURSOR is useful when more than one row of a table to be processed. It can
be loosely compared with sequential read of file. Usage of cursor involves four steps.
1.DECLARE statement.
This statement DECLARES the cursor. This is not an executable statement.
Syntax: DECLARE cursor-name CURSOR [WITH HOLD] FOR your-query
[FOR UPDATE OF column1 column2 | FOR FETCH ONLY]
2.OPEN statement.
This statement just readies the cursor for retrieval. If the query has ORDER
BY or GROUP BY clause, then result table is built. However it does not assign values
to the host variables.
Page:139
Syntax: OPEN cursor-name

3.FETCH statement.
It returns data from the results table one row at a time and assigns the value
to specific host variables.
Syntax: FETCH cursor-name INTO
:WS-Column1
:WS-Column2
The number of INTO variables should be equal to SELECT columns of your DECLARE
cursor statement. If FOR UPDATE OF clause is coded, then while fetching the row,
the exclusive lock is obtained over the page of the row. FETCH Cursor is usually
placed in the perform loop that will be executed until the SQLCODE is 100.SQLCODE
will be set to 100 when end of cursor is reached.
If FOR UPDATE OF clause is given, then all the columns that are going to be
updated using the cursor should be mentioned in this clause. If you dont mention
FOR UPDATE OF clause, you can update any number of columns using WHERE
CURRENT OF cursor-name but you will be getting the exclusive lock only during the
UPDATE statement and there are chances for the row to be locked exclusively by
another task in between your READ and UPDATE.
So it is suggested to use FOR UPDATE OF clause when you are using the
cursor for update purpose.
4.CLOSE statement.
It releases all resources used by the cursor. If you dont close the cursor,
COMMIT statement will close the cursor. To retain the cursor in OPEN stage during
COMMIT, Use WITH HOLD option in the DECLARE statement of the CURSOR. This
option will be effective only in batch environment. ROLLBACK close all the cursors
including the cursors defined with WITH HOLD option.
Syntax: CLOSE cursor-name.
Read Only Cursor
The cursor that cannot be used for any update purpose is called as READ
ONLY CURSOR. For example, cursor defined with UNION, JOIN or GROUP BY cannot
be used for update purpose.
DML-INSERT
INSERT statement is used to insert rows to the table. NOT NULL Column
values should be supplied during INSERT. Otherwise INSERT would fail.
1.Specify values for columns of single row to insert.
INSERT INTO YDEPT (DEPTNO, DEPTNAME, MGRNO, ADMRDEPT, LOCATION)
VALUES ('E31', 'DOCUMENTATION', '000010', 'E01', ' ');
INSERT INTO YEMP
VALUES ('000400', 'RUTHERFORD', 'B', 'HAYES', 'E31',
'5678', '1983-01-01', 'MANAGER', 16, 'M', '1943-07-10', 24000,
500, 1900);
Page:140
You can name all the columns or omit. When you list the column names, you
must specify their corresponding values in the same order as in the list of column
names.
2.Mass Insert. Another table or view contains the data for the new row(s).
INSERT INTO TELE
SELECT LASTNAME, FIRSTNME, PHONENO
FROM DSN8410.EMP
WHERE WORKDEPT = 'D21';
If the INSERT statement is successful, SQLERRD(3) is set to the number of
rows inserted.
Dependant table insertion: (Table with foreign Key)
Each non-null value you insert into a foreign key column must be equal to
some value in the primary key (the primary key is in the parent table). If any field in
the foreign key is null, the entire foreign key is considered null. If you drop the index
enforcing the primary key of the parent table, an INSERT into either the parent table
or dependent table fails.
DML-UPDATE
UPDATE statement is used to modify the data in a table. The SET clause
names the columns you want to update and provide the values you want them
changed to. The condition in the WHERE clause locates the row(s) to be updated. If
you omit the WHERE clause, DB2 updates every row in the table or view with the
values you supply.
If DB2 finds an error while executing your UPDATE statement (for instance, an
update value that is too large for the column), it stops updating and returns error
codes in the SQLCODE and SQLSTATE host variables and related fields in the SQLCA.
No rows in the table change (rows already changed, if any, are restored to their
previous values). If the UPDATE statement is successful, SQLERRD(3) is set to the
number of rows updated.
Syntax:
UPDATE table-name SET column1 =value1, column2 =value2 [WHERE condition];
DML-DELETE
DELETE statement is used to remove entire rows from a table. The DELETE
statement removes zero or more rows of a table, depending on how many rows
satisfy the search condition you specified in the WHERE clause. If you omit a WHERE
clause from a DELETE statement, DB2 removes all the rows from the table or view
you have named. The DELETE statement does not remove specific columns from the
row. SQLERRD(3) in the SQLCA contains the number of deleted rows.
Syntax: DELETE FROM table-name [WHERE CONDITION]
DML NULLS
One of the 12 CODD Rules for relation database system is NULL values are
supported for representing missing information in a systematic way irrespective of
the data type. DB2 supports NULL values.
NULL IN SELECT STATEMENT
NULL is defined as Unknown value in RDBMS terminology.
Two unknown values cannot be same. As EQUAL can be used only for known
values, COLUMN1=NULL is meaningless. In the where predicate if NULL needs to be
Page:141
checked, WHERE COLUMN1 IS NULL is the right way of coding. If GROUPBY is done
on a NULL column, then all the columns whose value is unknown (NULL) forms one
GROUP.
Example, If QUALIFICATION is a column defined with NULL attribute in
EMPTABLE, then SQL for retrieving all the employees whose QUALIFICATION is not
yet known is:
SELECT EMPNAME,QUALIFICATION FROM EMPTABLE
WHERE QUALIFICATION IS NULL.
High-level languages dont have any NULL concept. So if the column is NULL,
then the host variable corresponds to that column should be set to zero (numeric) or
spaces (non-numeric). This can be done in the programming. But for doing this we
should know whether the column is NULL or NOT. NULL indicators are used for this
purpose. NULL indicator is a 2 byte field (S9(4) COMP). Negative value,1 in NULL
indicator field indicate that the column is NULL.
EXEC SQL
SELECT QUALIFICATION
INTO :WS-QUALIFICATION :WS-QUALIFICATION-NULL
FROM EMPTABLE
WHERE EMPID=2052
END-EXEC
IF SQLCODE = 0
PERFORM 1000-CHECK-FOR-NULLS
.
100-CHECK-FOR-NULLS.
IF WS-QUALIFICATION-NULL < 0 THEN
MOVE SPACES TO WS-QUALIFICATION
END-IF.
If QUALIFICATION of EMPID = 2052 is NULL and if you didnt code null
indicator, then the query will fail with SQLCODE 305 and the error message
corresponding to this code is THE NULL VALUE CANNOT BE ASSIGNED TO OUTPUT
HOST VARIABLE NUMER posit-num BECAUSE NO INDICATOR VARIABLE IS
SPECIFIED
This kind of NULL check (100-CHECK-FOR-NULLS) is highly recommended for
numeric fields. Otherwise the program may abnormally end at some place down the
line, when the field is used for some arithmetic operation.
Instead of NULL indicator, VALUE function can also be used.
SELECT VALUE(QUALIFICATION, )
INTO :WS-QUALIFICATION
FROM EMPTABLE
WHERE EMPID=2052
If QUALIFICATION is NULL, then will be moved to WS-QUALIFICATION.
NULL IN INSERT/UPDATE STATEMENT
The concept is same. DB2 informs the NULL presence thru the NULL indicator
to the programming language. In similar way, the programming language should
inform the NULL to DB2 by NULL indicator.
Just before INSERT or UPDATE move 1 to NULL indicator variable and use
this variable along with host variable for the column to be inserted/updated.
Page:142
Once 1 is moved to null indicator, then independent of value presence in the host
variable, NULL will be loaded. In the following query, though B.E is moved to host
variable, it will not be loaded into the table as null indicator has 1. It should have
been set to 0 before loading.
MOVE 1 TO WS-QUALIFICATION-NULL.
MOVE B.E TO WS-QUALIFICATION.
EXEC SQL
UPDATE EMPTABLE
SET QUALIFICATION = :WS-QUALIFICATION :WS-QUALIFICATION-NULL
WHERE EMPID=2052
END-EXEC.
It is common feeling that NULL provides more trouble than benefit. So it is
always better to specify NOT NULL or NOT NULL WITH DEFAULT for all the columns.
Values in null indicator
A negative value in null indicator field implies that the column is NULL and a
positive value or ZERO in null indicator implies that the column is NOT NULL. Value of
2 in null indicator indicates that the column has been set to null as a result of a
data conversion error.
DCL-GRANT and REVOKE commands
Data Control Language consists of commands that control the user access to
the database objects. The Database Administrator (DBA) has the power to give
(GRANT) and take (REVOKE) privileges to a specific user, thus giving or denying
access to the data.
Syntax:
GRANT access ON object TO USER-ID | PUBLIC [WITH GRANT OPTION].
REVOKE access ON object FROM USER-ID | PUBLIC.
Example:
GRANT SELECT, UPDATE ON EMPTABLE TO SMSXL86.
GRANT ALL ON EMPTABLE TO SMSXL86.
REVOKE CREATE TABLE, CREATE VIEW FROM SMSXL86.
REVOKE INSERT ON EMTABLE FROM SMSXL86.
COMMIT and ROLLBACK

A transaction, or a unit of work, is a recoverable sequence of one or more
SQL operations that are grouped together as a single unit, usually within in an
application process.
Page:143
All the changes made to the database since the initiation of transaction, are
made permanent by COMMIT. ROLLBACK brings back the database to the last
committed point.
Syntax:
EXEC SQL COMMIT END-EXEC
EXEC SQL ROLLBACK END-EXEC
DB2 Restart Logic:
Usually there is only one COMMIT or ROLLBACK just before the termination of
the transaction. But it is not preferred always.
If the program is updating one million records in the table space and it
abnormally ends after processing ninety thousand records, then the issued
ROLLBACK brought the database back to the point of transaction initiation.
The restart of the program must have to process ninety thousand successful-but-notcommitted updates once again. The repeated cost occurred is due to Bad design of
the application.
If the program is expected to do huge updates, then commit frequency has to
be chosen properly. Let us say, after careful analysis, we have designed our
application COMMIT frequency as thousand records. If the program abnormally ends
while processing 1500th record, then the restart should not start from first record but
from 1001st record. This is done using restart logic.
Create one temporary table called RESTARTS with a dummy record and
inserts one record into the table for every commit with key and occurrence of
commit. This insertion should happen, just BEFORE the issue of COMMIT.
First paragraph of the procedure should read the last record of the table and
skipped the records that are already processed and committed (1000 in the previous
case). After the processing of all the records (one million), delete the entries in the
table and issue one final COMMIT.
JOINS
Most of the times, the complete information about an entity is retrieved from
more than one table. Retrieving information from more than one table can be done
Page:144
in two ways. One method is JOIN and the other method is UNION. We have already
discussed about UNION. It is used to get information about more entities. In other
words, it returns more rows. JOIN is used to get detail information about entities. In
other words, it returns more columns.
There are two kinds of JOIN. Inner join returns rows from two or more tables
based on matching values. Outer join returns rows from tow or more tables
independent of matching or non-matching.
JOIN: INNER
Two or more tables are joined together using the column having equal values
among them.
EMPTABLE
EMP_ID EMP_NAME DESIG
100MUTHU
TL
101DEVI
SSE
DESIG
SALARY
SSE
400000
SE
300000
SALTABLE
SELECT EMP-NAME SALARY FROM EMPTABLE, SALTABLE WHERE

EMPTABLE.DESIGNATION = SALTABLE.DESIGNATION
It can also be coded as
SELECT EMP-NAME SALARY FROM EMPTABLE INNER JOIN SALTABLE ON
Result
DEVI
400000
Since there is no TL designation in the SALTABLE, MUTHU will not appear in the
output.
JOIN:FULL OUTER JOIN
The clause FULL OUTER JOIN includes unmatched rows from both tables.
Missing values in a row of the result table contain nulls.
SELECT EMP-NAME SALARY FROM EMPTABLE FULL OUTER JOIN SALTABLE ON
Result:
MUTHU -DEVI
400000
--300000
JOIN:LEFT OUTER JOIN
The clause LEFT OUTER JOIN includes rows from the table named before it
where the values in the joined columns are not matched by values in the
joined columns of the table named after it.
SELECT EMP-NAME SALARY FROM EMPTABLE LEFT OUTER JOIN SALTABLE ON
Result
MUTHU --
Page:145
DEVI
400000
JOIN:RIGHT OUTER JOIN
The clause RIGHT OUTER JOIN includes rows from the table named after it
where the values in the joined columns are not matched by values in the
joined columns of the table named before it.
SELECT EMP-NAME SALARY FROM EMPTABLE RIGHT OUTER JOIN SALTABLE ON
Result
DEVI
----
400000
300000
Restrictions on Outer Joins

1. VALUE or COALESCE function can be used only with FULL OUTER join.
2. Only = can be used as comparison operator in FULL OUTER join.
3. Multiple conditions can be coded with AND. OR and NOT are not allowed.
Sub-queries
Sub-Queries are nested SELECT statements. Sub-query enables a user to
base the search criteria of one SELECT on the result of another SELECT statement.
It is always part of a predicate.
Syntax: (MAIN-QUERY) WHERE operand operator (SUB-QUERY)
Sub-queries can appear in the predicates of other sub-query. This is called
nested sub-query and the nesting is possible up to 15 levels.
Usually sub-query executes only once and it is executed first and the value
returned by sub-query will be used for main query. This kind of sub-query is called as
un-correlated sub-query. When the sub-query refers to the table referred in the
main-query then the sub-query is executed for every execution of main-query and
this kind of query is called correlated sub-query. Correlated sub-queries cannot be
used for update/insert/delete purpose.
Example:
The following query lists the employees who get more than the average salary
of their department.
SELECT EMPNAME,DEPTNUM,SALARY
FROM EMPTABLE
WHERE SALARY > (SELECT AVG(SALARY)
FROM DEPTNUM
WHERE EMPTABLE.DEPTNUM=DEPT.DEPTNUM)
OPTIMIZER- ACCESS PATH SELECTION
Generally, the fastest way to access DB2 data is with index. Indexes are
structured in such a way to increase the efficiency of finding a particular piece of
data. DB2 uses many different algorithms to traverse an index structure. These
algorithms are designed to elicit optimum performance in a wide variety of data
access scenarios.
Page:146
To use an index:
1. One of the predicates for the SQL statement must be index-able.
2. One of the columns (in any index-able predicate) must exist as column in an
available index.
There
1.
2.
3.
4.
5.
6.
are various types of indexes are available. They are:

Direct Index Look-Up.
Index Scan Matching index scan.
Index Scan Non-matching index scan.
Index-Only access.
List pre-fetch
Multi-index access (AND & OR)
Example: The primary key of the table TASKTAB is a composite key with columns
CLIENT_ID, PROJECT_ID, MODULE_ID.
Direct Index Look-Up:
Values are provided for each column in the index. In the above example, if
the value for all the three columns are given in the where clause, then there will be
a direct index look up.
SELECT MODULE_ID, MODULE_STATUS FROM TASKTAB
WHERE CLIENT_ID = 100 AND PROJECT_ID=10 AND MODULE_ID = 1
If only CLIENT_ID and MODULE_ID is given then direct index look-up is not possible.
Index scan would occur.
Matching Index Scan or Absolute Positioning
It starts with root page and work down to a leaf page in much the same
manner as direct index loop up does. However since the full key is unavailable, DB2
must scan the leaf pages using the values it does have, until all matching values
have been retrieved. As it starts at root page, the value of first column(s) should be
given.
SELECT MODULE_ID, MODULE_STATUS FROM TASKTAB
WHERE CLIENT_ID = 100 AND PROJECT_ID=10
The query use matching index scan and returns all the module id and module
status for the given client and project.
Non-matching index scan or Relative positioning
Non-matching index begins with the first leaf page in the index and
sequentially scans subsequent leaf pages applying the available predicates. If any
ORDER BY or GROUP BY is coded, then optimizer prefers non-matching index scan
than table space scan.
SELECT CLIENT_ID, MODULE_ID, MODULE_STATUS FROM TASKTAB
WHERE PROJECT_ID=10
ORDER BY MODULE_ID
Index-only access
If all the data required by the query is available in index, then DB2 can avoid
reading the data page and perform non-matching index scan.
This special type of non-matching index scan is called index-only access.
Page:147
SELECT CLIENT_ID FROM TASKTAB WHERE PROJECT_ID = 10 AND MODULE_ID=1

Sequential pre-fetch and List pre-fetch
Sequential pre-fetch is a technique used by optimizer during SEQUENTIAL
table space SCAN. It can be though of as a read ahead mechanism invoked to pre-fill
DB2 buffers so that data is already in memory before it is requested.
List Pre-fetch is a technique used by optimizer during INDEXED access. If the
cluster ratio of the index is poor (less than 80%) then RIDs of matching indexes are
sorted based on data page number and then the data-page is accessed.
This significantly reduces the I-O as same page will not be read second time
in this method. Clustered index access doesnt use this technique as already the data
rows are in the same order of index. (Cluster ratio is 1)
Multi index access
When more than two indexes are used in where clause..
1.Using the first index, RIDs matching the requirement are chosen and sorted based
on data-page number.
2.Using the second index, RIDs matching the requirement are chosen and sorted
based on data-page number.
3.If the indexes are joined using OR, then the union of RID list of 1 and 2 is used to
access data page.
4. If the indexes are joined using AND, then the intersection of RID list of 1 and 2
is used to access data page.
Table Space Scan
If there is no index-able column in the predicate or DB2 decides not to use
available index, then the optimizer choose table space scanning.
Table space scanning is like sequential read of file. All the rows of all the
pages are read sequentially and the rows satisfying the selection criteria in the where
clause is chosen for further processing.
Query Parallelism
This is another technique that can be applied by the optimizer is query
parallelism. When parallelism is invoked, an access path can be broken up into
parallel groups. Each parallel group represents a series of concurrent operations with
the same degree of parallelism. Degree of parallelism refers to the number of
concurrent tasks used to satisfy the query.
Query I/O parallelism (as of DB2 V3):
Breaks the data access for the query into concurrent I/O streams executed in
parallel should reduce the overall elapsed time for the query. With query I/O
parallelism, DB2 is limited to operate on a single processor for each query.
Query CP Parallelism (as of DB2 V4):
The large query is decomposed into multiple smaller queries that can be
executed concurrently with one another on multiple processors. CP parallelism
always uses I/O parallelism. This should further reduce the elapsed time of the
query.
Query Sysplex Parallelism (as of DB2 V5):
Page:148
It further enhances parallel operations by enabling a single query to be

broken up and run across multiple DB2 subsystems within a data-sharing group. It
takes the advantage of the processing power of multiple DB2 subsystems.
Filter factor
Filter factor is a ratio that estimates I/O costs. The optimizer calculates the
filter factor for a querys predicates based on the number of rows that will be filtered
out by the predicates. That is, if the predicate is most restrictive, the filter factor
would be low, the I-O cost would be less and the query will be more efficient.
For an indexed column, FIRSTKEYCARDF column of SYSIBM.SYSINDEXES consists of
number of distinct values in the column.
If the predicate is column=value then filterfactor = 1/FIRSTKEYCARDF
If the predicate is column <> value then filterfactor = (1-1/ FIRSTKEYCARDF).
This case involves more I-O is involved and inefficient.
EXPLAIN
We have already seen that DB2 optimizer chooses the access path for the
query. If we want to know what access path DB2 chose during the plan preparation,
then the request should be placed to DB2. The request can be done two ways:
Method 1: Use EXPLAIN (YES) parameter in the BIND card.
SELECT * FROM ownerid.PLAN_TABLE ORDER BY APPLNAME, COLLID, VERSION,
PROGNAME, TIMESTAMP DESC, QUERYNO, QBLOCKNO, PLANNO
Method 2: Use the following command directly in the program or in SPUFI or QMF.
Syntax: EXPLAIN ALL SET QUERYNO=integer FOR SQL-statement.
Before posting the request, there should be a PLAN_TABLE under the user-id,
based on model SYSIBM.PLAN_TABLE. PLAN_TABLE is a standard table that must be
defined with predetermined columns, data types and lengths. During bind process,
DB2 optimizer briefs the access path chose by it in the PLAN_TABLE.
If you want to query the access path for single query then use the query
below:
SELECT * FROM PLAN_TABLE WHERE QUERYNO=integer OREDERBY QBLOCKNO,
PLANNO.
Now the access path taken by the optimizer is loaded into PLAN_TABLE. One
should know the meaning of PLAN_TABLE columns and values to understand the
access path. DB2V5 has 46 columns in PLAN_TABLE.
PLAN_TABLE Column name
QUERYNO
QBLOCKNO
APPLNAME
PROGNAME
Meaning
Integer value assigned by the user issuing the EXPLAIN
or DB2
Integer identifying level of sub-query or union in a SQL
statement. The first sub-select is numbered as 1, the
second as 2 and so on.
Name of the PLAN or PACKAGE based on BIND stage of
static SQL. It will be spaces for SPUFI request.
Name of the program in which SQL statement is
embedded. DSQIESQL for SPUFI EXPLAIN
PLANNO
METHOD
ACCESSTYPE
MATCHCOLS
INDEXONLY
PREFETCH
PARALLELISM MODE
QBLOCK_TYPE
JOIN_TYPE
TIMESTAMP
Page:149
Integer value saying the step of the plan in which

QBLOCKNO is processed.
(The order in which plan steps are undertaken)
Integer value identifying the access method used for
the given step.
1- Nested loop join
2-Merge scan join
3-independent sort as a result of ORDERBY, GROUP BY,
DISTINC or IN
4-Hybrid Join.
Technique used for access.
I Indexed access
I1 One fetch index scan
R Table Space Scan
N Index access with In predicate
M Multiple index scan
MX specification of the index name for multiple index
access.
MI Multiple index access by RID intersection
MU Multiple index access by RID union
Number of index columns used in indexed access.
Access type should be I, I1, M, MX.
Y indicates access to the index is sufficient to satisfy
the query.
Indicates which type of pre fetch will be used.
SSequential, L- List, BlankUnknown at bind time.
Indicates type of parallelism used at bind time.
I I/O Parallelism, C CPU Parallelism
X Sysplex Parallelism, Blank No or determined at
run time.
Type of SQL operation performed for the query block.
CORSUB Correlated sub query
NCOSUB- Non correlated sub query
UPDCUR, DELCUR Update and Delete using WHERE
CURRENT OF option of cursor.
SELECT, INSERT, UPDATE, DELETE Type of DML.
SELUPD SELECT with FOR UPDATE OF.
Indicates type of join implemented.
F Full Outer Join
L Left Outer Join (or converted right outer join)
Blank Inner join or no join.
Date and time the EXPLAIN for this row was issued.
Other columns are:

CREATOR, TNAME, TABNO, ACCESSCREATOR, ACCESSNAME, SORTN_UNIQ,
SORTN_JOIN, SORTN_ORDERBY, SORTM_GROUPBY, SORTC_UNIQ, SORTC_JOIN,
SORTC_GROUPBY, SORTC_GROUP, TSLOCKMODE, REMARKS, COLUMN_FN_EVAL,
MIXOPSEQ, VERSION, COLLID, ACCESS_DEGREE, ACCESS_PGROUP_ID,
JOIN_DEGREE, JOIN_PGROUP_ID, SORTC_PGROUP,ID, SORTN_PGROUP_ID,
MERGE_JOIN_COLS, CORRELATION_NAME, PAGE_RANGE, GROUP_MEMBER,
IBM_SERVICE_DATA,WHEN_OPTIMIZE, BIND_TIME.
Page:150
DYNAMIC SQL
Static SQL is embedded in application program and the only values that can
change are the values of the host variables in the predicates. Dynamic SQL are
characterized by the capability to change columns, tables and predicates during a
programs execution.
Advantages: Flexibility and best access path (as the bind happens at the run
time using latest RUNSTATS information)
Disadvantage: Slow as the runtime is bind time + execution time.
Types of Dynamic SQL
1.EXECUTE IMMEDIATE:
Move the SQL statement to be executed into the host variable and issue
execute immediate command.
01 WS-HOST.
49 WS-HOST-LENGTH PIC S9(04) COMP.
49 WS-HOST-VAR
PIC X(40).
MOVE +40 TO WS-HOST-LENGTH
MOVE SET EMP_NAME = MUTHU WHERE EMP_NO=2052 TO WS-HOST-TEXT.
EXEC SQL EXECUTE IMMEDIATE: WS-HOST-VAR END-EXEC.
Disadvantages:
1.It cannot be used for SELECT.
2.Executable form of the SQL will be deleted once it is executed.
2.EXECUTE WITH PREPARE:
This is equivalent to first form but here the SQL is prepared before execution
and so the second disadvantage is avoided in this method.
Form: 1
MOVE SET EMP_NAME = MUTHU WHERE EMP_NO=2052
EXEC SQL PREPARE RUNFORM FROM :WS-HOST END-EXEC.
EXEC SQL EXECUTE RUNFORM END-EXEC.
TO WS-HOST-TEXT.
Form: 2
Parameter markers can be used in place of constant values. This acts like
placeholder for the host variables in the SQL.
MOVE SET EMP_NAME = MUTHU WHERE EMP_NO=?
TO WS-HOST-TEXT.
EXEC SQL PREPARE RUNFORM FROM :WS-HOST END-EXEC.
MOVE 2052 TO WS-EMP-NUM
EXEC SQL EXECUTE RUNFORM USING :WS-EMP-NUM END-EXEC.
Disadvantage:
Select statement is not supported.
Page:151
3.SELECT with STATIC Columns (Fixed-list SELECT)

The following method is used for data retrieval but the column to be selected
are known and unchanging.
Let the query be SELECT EMP_NAME FROM EMPTABLE WHERE DEPT_NO=?
MOVE Select SQL to WS-HOST.
EXEC SQL DECLARE CSR1 CURSOR FOR SELSQL END-EXEC
EXEC SQL PREPARE FROM :WS-HOST END-EXEC
MOVE 100 TO WS-DEPT-NO
EXEC SQL OPEN CSR1 USING :WS-EMP-NO END-EXEC
PERFORM UNTIL SQLCODE =100
EXEC SQL FETCH CSR1 INTO :WS-EMPNAME END-EXEC
END-PERFORM.
EXEC SQL CLOSE CSR1 END-EXEC.
Disadvantage:
1.Table name cannot be changed. Number of columns cannot be modified during run
time.
The fourth type of dynamic SQL known as varying-list select provide hundred
percent flexibility where the number of columns and even the table we are accessing
can be changed during run time. This uses the copybook SQLDA for its purpose.
INTEGRITY
Integrity means the accuracy, correctness or validity of data contained in the
database. Maintaining integrity is not an easy task in a multi user environment. So
the task of maintaining integrity is handled by the system than the user.
Domain Integrity.
It ensures that a value in a column is a member of columns domain or legal
set of values. Simple example is, alphabetic data on the integer column is restricted.
Entity Integrity.
It means every row in a table is unique. To ensure entity integrity, developer
has to define a set of column(s) as primary key.
Referential Integrity.
It ensures the data integrity between the tables related by Primary (Parent)
and Foreign (Child) keys.
Page:152
Concurrency and its issues:

DB2 allows more than one application to access the same data at essentially the
same time. This is known as concurrency. Concurrency results the following
problems.
1.The LOST UPDATE Problem.
Time
T0
T1
T2
T3
Transaction-A
Read row R1.
Transaction-B
Read row R1.
Update row R1.
Update row R1.

Transaction-B overwrites the change done by Transaction-a at T2.
In other words, the update of transaction-A is lost.
2.Uncommitted Dependency Problem
It arises when a transaction is allowed to retrieve (update) a row that has
been updated by another transaction and has not yet been committed by the other
transaction.
Time
Transaction-A
Transaction-B
T0
Update row R1.
T1
Read and Update row R1.
T2
Rollback
Transaction-B updates row R1 based on the uncommitted change done by
Transaction-A. Transaction-B rolled back its changes.
3.Data inconsistency.
Let us say there is 2 account A1,A2 and account A3 should have sum of
amounts in A1 and A2. Say A1 = 1000 A2 = 2000 A3=3000
Time
Transaction-A
Transaction-B
T0
Read row A1.(1000)
T1
Read row A1 and update to 2000.
T2
Update row A2.(3000)
T3
Calculate A3 and update. (4000)
T4
Commit
Commit
Now the value of A1=2000, A2=3000 but A3=4000 == > Data inconsistent
as A3 is expected to be 5000 as per design.
Page:153
Locking
Locking solves the concurrency issues described above. Locking prevent
another transaction to access the data that is being changed by the current
transaction. Three main attributes of lock are mode, size and duration.
Mode of the lock:
It specifies what access to the locked object is permitted to the lock owner
and to any concurrent application process.
Exclusive (X) The lock owner can read or change the locked page. Concurrent
processes cannot acquire ANY lock on the page nor they access the locked page.
Update (U) The lock owner can read but cannot change the locked page. However
the owner can promote the lock to X and update. Processes concurrent with the U
lock can acquire S lock and read the page but another U lock is not possible.
Share (S) - The lock owner and any concurrent processes can read but not change
the locked page. Concurrent processes can acquire S or U locks on the page.
The above three are with respect to pages. Three more kind of locks are possible at
table /table space level. They are: Intent Share (IS), Intent Exclusive (IX) and
Share/Intent Exclusive (SIX)
SIZE of the lock
It can be ROW, PAGE, TABLE and TABLESIZE.
Duration of the lock
It is the length of time the lock is held. It varies according to when the lock is
acquired and released. This is controlled by ACQUIRE, RELEASE and ISOLATION
LEVEL parameters of BIND.
LOCK Issues
Locking is introduced to avoid concurrency issues. Locking resulted suspension,
timeout and deadlock problems. But LOCK is needed for data integrity.
Suspension.
IRLM suspends an application process if it requests a lock on an object that is
already owned by another application process and cannot be shared. The suspended
process is said to be in Wait Stage for the lock. The process resumes when the lock
is available.
Timeout.
Page:154
When the object is in wait stage more than predefined time, then the process
(program) is terminated by DB2 with SQL return code 911 or 913 meant for
TIMEOUT. The IRLMRWT of DSNZPARM determines the duration of the wait time.
Deadlock.
When two or more transactions are in simultaneous wait stage, each waiting
for one of others to release a lock before it can proceed, DEADLOCK occurs.
If deadlock occurs, transaction manager of DB2 breaks it by terminating one
of the transactions and allows the other to process. Later, the application
programmer will restart the terminated transaction.
Lock and Latch
A true lock is handled by DB2 using an external component called IRLM.
However, when it is practical, DB2 tries to lock pages without going to IRLM.
This type of lock is known as LATCH. Latches are internally set by DB2 without going
to IRLM.
Initially latches are used to lock only DB2 index pages and internal DB2
resources. DB2 V3 uses latching more frequently for data pages also. Latches are
preferred when a resource serialization is required for a short time. Locks and
Latches guarantee data integrity at the same level.
Advantage of Latch over Lock:
1.Cross-memory service calls to IRLM is eliminated.
2.Latch requires about one third the number of instructions as a lock.
Lock Promotion
When lock is promoted from lower level to upper level, it is called lock
promotion. DB2 acquire U lock to update a row. But at the same concurrent
processes can access the row in S lock. During update, it needs X lock and to get
the X lock, it waits for concurrent processes to release S locks. It promotes the
current lock from U to X and this is called LOCK PROMOTION ON MODE. (U
Update in Pending and X Update in Progress)
When the number of page locks exceeds predefined limit, it is promoted to
table or table space level and this is LOCK PROMOTION ON SIZE.
Lock Strategy
DB2 decides the lock to be applied based on the following:
1. Lock size declared at the time of table space creation.
2. Type of SQL statement SELECT or UPDATE
3. Explicit LOCK statement in the program.
4. ACQUIRE and RELEASE option chosen by user at the BIND time.
5. Isolation level specified at bind time.
6. The NUMLKTS and NULLKUS limit.
Page:155
Normalization
Data Normalization describes the process of designing a database and
organizing data to take best advantage of relational database principles. It is the
process of putting one fact in one appropriate place. This optimizes the updates at
the expense of retrievals.
When a fact is stored in only one place, retrieving many different but related
facts usually requires going to many different places. This tends to slow the retrieval
process but update is easy and faster as the fact you are updating exists in only one
place.
The process involves five levels. Tables are usually normalized till level three.
Un normalized Data
Columns should be automatic and

Remove repeating groups
First Normal Form
Remove dependence of non-key
Column(s) on part of the key
(Partial Dependence)
Second Normal Form
Remove dependence of non-key
Column(s) on other non-key columns
(Transitive Dependence)
Third Normal Form
The three levels can be memorized using the statement below:
Page:156
The values in a row are dependant on the key, the whole key, and nothing
but the key, so help me CODD
In the fourth level, multi-valued dependencies are removed and in the fifth level,
remaining anomalies are removed.
De-normalization
De-normalization is the process of putting one fact in more than one place. It
is the reverse process of normalization. This will speed up the retrieval process at the
expense of data modification. De-normalization is not a bad decision when a
completely normalized design is not performing optimally.
System Catalog Tables

DB2 catalog is contained in a single database (DSNDB06). The 54 tables in
the DB2 catalog collectively describe the objects and resources available to DB2.
Prior to Version5, there was an additional database called Communication
Database (CDB) was used for establishing and documenting distributed DB2
connections. In the version 5, it has been renamed and rolled into DB2 catalog.
Each and every table has its own importance. So explaining all the tables and
the columns in it is highly impossible. Refer the query section to know about some of
the frequently used queries on SYSIBM tables.
1. SYSIBM.SYSDUMMY1 is a dummy table. When you want to do some arithmetic
operation or retrieving system information like current-date, you can query this
table.
SELECT 10.25*20.25 FROM SYSIBM.SYSDUMMY1;
2. When a DBRM is bound to PLAN, all the SQL statements are placed into
SYSSTMTDB2 table. When a DBRM is bound to PACKAGE, all of its SQL
statements are placed into SYSPACKSTMT table. Information about the DBRMS
that were bound to PLAN or PACKAGE, are stored in SYSDBRM table. DBRM just
created but not yet bound is not registered anywhere. (Remember, pre-compiler
dont use DB2 at all and DBRM is an output of pre-compiler)
3. Information about plan is stored in SYSDBRM, SYSPACKAUTH, SYSPACKLIST,
SYSPLAN, SYSPLANDEP, SYSPLANAUTH, SYSPLSYSTEM, SYSSTMT and
SYSTABAUTH.
4. Information about the package is stored in SYSPACKAGE, SYSPACKAUTH,
SYSPACKDEP, SYSPLSYSTEM, SYSSTMT and SYSTBAUTH.
5. DB2 Catalog stores only information about the PLAN. The executable form of
PLAN, called Skeleton cursor table (SKCT), is stored in the DB2 directory in the
SYSIBM.SCT02 table.
Page:157
6. Information about Table space/ Table and Indexes are stored in the following
Tables.
SYSTABLESPACE
SYSTABLES
SYSCOLUMNS
SYSRELS
SYSINDEXES
SYSTABLEPART
SYSCOLUMNS
SYSTBAUTH
SYSCOLAUTH
SYSFIELDS
SYSLINKS
SYSINDEXPART
SYSCOLDIST
SYSFOREIGNKEYS
SYSKEYS
SYSCOLDISTSTATS
SYSTABSTATS
SYSINEXSTATS
Triggers (Available in version 6)
Trigger is a piece of code that is executed in response to a data modification
statement. (insert/update/delete). To be a bit more precise: triggers are eventdriven specialized procedures that are stored in, and managed by the RDBMS. Each
trigger is attached to a single, specified table. Triggers can be thought of as an
advanced form of "rule" or "constraint" written using an extended form of SQL.
Therefore triggers are automatic, implicit, and non-bypass able.
Nested Triggers
A trigger can also contain insert, update, and delete logic within itself.
Therefore, a trigger is fired by a data modification, but can also cause another data
modification, thereby firing yet another trigger. When a trigger contains insert,
update, and/or delete logic, the trigger is said to be a nested trigger. Most DBMSs,
however, place a limit on the number of nested triggers that can be executed within
a single firing event. If this were not done, it could be quite possible to having
triggers firing triggers ad infinitum until all of the data was removed from an entire
database!
Triggers cannot be attached to the following tables:
A System Catalog Table, PLAN_TABLE, STATEMENT_TABLE, DSN_FUNCTION_TABLE
View, Alias, Synonym, Any table with a three-part name
Trigger in place of RI
Triggers can be coded, in lieu of declarative RI, to support ALL of the RI rules.
We can specify only ON DELETE rules in the DDL statement. UPDATE and INSERT
rules of RI can be implemented using triggers.
Sample Trigger:
CREATE TRIGGER SALARY_UPDATE
BEFORE UPDATE OF SALARY
ON EMP
FOR EACH ROW MODE DB2SQL
Page:158
WHEN (NEW.SALARY > (OLD.SALARY * 1.5))

BEGIN ATOMIC
SIGNAL SQLSTATE '75001' ('Raise exceeds 50%');
END;
The trigger executes once for each row. So if multiple rows are modified by a
single update, the trigger will run multiple times, once for each row modified. Also,
the trigger will be run BEFORE the actual modification occurs. Finally, take special
notice how NEW and OLD are used to check values before and after the update.
Triggers Vs Stored Procedures
Triggers are also similar to stored procedures. Both consist of procedural logic
that is stored at the database level. However, stored procedures are not event-driven
and are not attached to a specific table. A stored procedure is explicitly executed by
invoking a CALL to the procedure (instead of implicitly being executed like triggers).
Additionally, a stored procedure can access many tables without being specifically
associated to any of them. DB2 has supported stored procedures since Version 4.
DB2 Utilities
Most of the time, application programmer doesnt need any knowledge on
DB2 utilities. Backup, Reorganization, Recovery and maintenance of database are
taken care by the DBA of the project.
DB2 utilities are briefed here to just familiar with it.
DSNUPROC is the DB2 catalogued procedure that executes the program DSNUTILB.
DSNUTILB is a master program and it calls other programs based on the first word of
control card. The first word of control card should be utility that needs to be invoked.
Most of the DB2 Utility programs are divided into four major categories.
Utility
COPY
(Backup and
Recovery)
MERGECOPY
(Backup and
Recovery)
QUIESCE
(Backup and
Recovery)
RECOVER
(Backup and
Recovery)
LOAD
Purpose
It is used to create image copy backup dataset for a complete
table space or a single partition of a table space. It can be of
type FULL or incremental. Based on number of modified pages
after the previous backup, prefer FULL or incremental image
copy. Successful copy details are loaded in SYSIBM.SYSCOPY.
It is used to create full image copy from incremental image
copies.
It is used to record a point of consistency for related
application or system tables. Usually done before image copy.
It is used to restore DB2 table space and indexes to a specific
point in time. It uses image copy and DB2 logs information to
roll back the table space content to the old one. RECOVER
INDEX generated new index data from current table space
data. It is actually regeneration than restoring. So it is
followed by RECOVERY TABLESPACE, which is restoring.
It is used to accomplish bulk inserts to DB2 tables. It can add
Page:159
(Data
Organization)
rows to a table retaining the current data or it can replace the

existing data with new data.
REORG
(Data
Organization)
It is used to re-clustering the data in table space and resetting

the free space to the amount specified in the CREATE DDL and
deletes and redefining VSAM datasets for STOGROUP defined
objects.
These are data consistency utilities. They are used to monitor,
control and administer data consistency errors.
CHECK, REPAIR,
REPORT,
DIAGNOSE
RUNSTATS
(Catalog
Manipulation
Utility)
It collects statistical information for DB2 tables, table spaces,

partitions, indexes and columns. It places this information into
DB2 catalog tables. The DB2 optimizer uses the data in these
tables to determine optimal access path for the SQL queries.
CATMAINT, MODIFY and STOSPACE are the other catalog
manipulation utilities.
How the load card of LOAD utility looks?

LOAD DATA INDDN(SYSREC) LOG(NO) NOCOPYPEND
RESUME YES
=> Resumes all the records from first to last.
ENFORCE CONSTRAINTS
INTO TABLE DB1.EMPTABLE
(EMPNAME POSITION(1) CHAR(20),
EMPADDRESS POSITION(25) VARCHAR
EMPPROJECT POSITION(50) CHAR(3))
RESUME YES => Resumes all the records from first to last.
RESUME NO => Table should be empty.
REPLACE
=> It overrides the previous record set with current record set.
SYSREC dataset should contain
1
XXXX
YYYYY
25
XXXXXXXXXXXXXXXXXXXX
YYYYYYYYYYY
50
XXXXX
YYYYY
LOAD Data statement describes the data to be loaded. In the above example,
SYSREC is coded and it should be one of the DDNAME in JCL, which points to actual
dataset.
Data are loaded in the order they appear in the dataset. Automatic data conversion
between compatible data types is done if needed.
What are COPY PENDING and CHECK PENDING status?
COPY PENDING and CHECK PENDING status on the table space restricts any
updates on table space.
If you LOAD or REORG the table space with the option LOG NO, then the table
space get into COPY PENDING status. The meaning is, an image copy is needed on
the table space.
Page:160
If you LOAD the table space with ENFORCE NO option, then the table space
get into CHECK PENDING status. The meaning is table space is loaded without
enforcing constraints. CHECK utility needs to be run on the table space.
COPY Utility
COPY TABLESPACE EDSDB.TRG007TS
FULL YES |NO
YES FOR FULL IMAGE and NO FOR INCREMENTAL
SHRLEVEL CHANGE
Page:161
MERGECOPY Utility
MERGECOPY TABLESPACE EDSDB.TRG007TS
NEWCOPY (YES/ NO)
YES. Merging all the incremental copies with the available full image copy and
creates a new image copy.
NO. Merging all the incremental copies and prepares a single incremental image
copy.
REPAIR Utility
//DSNUPROC.SYSIN *
REPAIR OBJECT LOG YES SET TABLESPACE NPCBTC00.TASSEM01 NOCOPYPEND
/*
This card repairs the table space NPCBTC00.TASSEM01 and clears the copy pending
status on the table space.
RUNSTATS Utility
//SYSIN DD *
RUNSTATS TABLESPACE NPCBTC00.TASSEM01
TABLE (NPCBT9TS.TASTEM01)
COLUMN (NPA, TELNO, CLS_MO, TR_CRT_MO, TR_CRT_YY,
SVC1_TYP, RPT1_TYP, SRC1_TYP, CLR_MO, CLR_DY)
INDEX (ALL)
SHRLEVEL CHANGE
REPORT YES UPDATE ALL
/*
REBUILD/RECOVER/REORG and CHECK DATA Utility
REBUILD INDEX (ALL) TABLESPACE(NPCBT9DB.NPCS7011)
SORTKEYS
RECOVER TABLESPACE NPCTT9DB.NPCS9042
REORG TABLESPACE NPCTT9DB.NPCS7048
UNLDDN SYSRE
SORTDEVT SYSDA
SORTNUM 4
WORKDDN(SYSUT1,SRTOUT)
CHECK DATA TABLESPACE NPCTT9DB.NPCS7013
SORTDEVT SYSDA
SORTNUM 4
WORKDDN(SYSUT1,SRTOUT)
QUIESCE Card:
QUIESCE
TABLESPACE NPCBT900.NPCS1002
TABLESPACE NPCBT900.NPCS1004
Page:162
START Database:
DSN SYSTEM(DB2X)
START DATABASE(NPCBT900) SPACENAM(NPCS1001) ACCESS(RO)
START DATABASE(NPCBT900) SPACENAM(NPCS1002) ACCESS(RO)
END
SQL Guidelines for better performance
1.
2.
3.
4.
5.
Pre-test all your queries in SPUFI or QMF before placing in program.

Avoid qualifiers for the tables in the program. Give it on the BIND card.
Unless really needed, avoid usage of DISTINCT, UNION and ORDER BY.
Avoid WHENEVER.
Code predicates on index-able columns Code the most restrictive predicated
first Use equivalent data types in the predicates
6. Do not use arithmetic expressions in a predicate. DB2 will not use index in
this case.
7. Prefer BETWEEN in stead of <= and >=
8. Prefer IN in stead of LIKE
9. Avoid the usage NOT. (Except with NOT EXISTS)
10. Join using SQL instead of programming logic.
11. Use JOINS than Sub-query.
12. Minimize the number of tables in the JOIN.
13. When a SELECT statement is coded, use FOR FETCH ONLY. This enables block
fetch option DB2 and leads to efficient retrieval.
14. If the cursor is used for update, use FOR UPDATE OF clause in DECLARE
statement itself.
15. Though DECLARE CURSOR can be given in procedure division, code it in
working storage section as a good practice. DECLARE cursor is not an
executable statement and procedure division is for executable statements.
16. For existence checking, use constant. (..WHERE EXISTS (SELECT 1 FROM
TAB1)
17. Give a thought on dropping indexes before large insertions. It avoids
unorganized indexes and poor performance.
18. Dont USE Select * - If there is any change in DB2 table structure, then the
program needs to be modified though it doesnt need the field.
There is unnecessary I-O overhead.
19. Do the arithmetic operations while selecting the column(s) than doing in
programming language
Famous SQL Queries

First and second maximum salary of EMPTABLE
SELECT MAX(SALARY) FROM EMPTABLE
SELECT MAX(SALARY) FROM EMPTABLE
WHERE SALARY < (SELECT MAX(SALARY) FROM EMPTABLE)
Nth maximum salary from EMPTABLE
SELECT SALARY FROM EMPTABLE A
WHERE N = (SELECT COUNT(*)FROM EMPTABLE B
WHERE B.SALARY >= A.SALARY)
How do you calculate day of the week?
SELECT DAYS(DATE(:WS-SQL-DT1))-(DAYS(DATE(:WS-SQL-DT1))/7) * 7
INTO :WS-DAY-OF-WEEK
FROM SYSIBM.SYSDUMMY1
0-6 means SUNDAY to SATURDAY respectively.
How do you get current date in DB2?
SET :WS-CURRENT-DATE = CURRENT DATE
Sample query to concatenate two columns
SELECT Mr. ||FIRST_NAME||STF_LAST_NAME
INTO :WS-USER-NM
FROM CRS_STAFF
WHERE STF_CODE = :WS-I-USER-ID
How do I get the primary key of a table?
SELECT NAME
FROM SYSIBM.SYSCOLUMNS
WHERE TBCREATOR = <TABLE CREATOR NAME>
AND TBNAME = <TABLE NAME>
AND KEYSEQ > 0
Or
SELECT A.COLNAME,
A.COLSEQ,
A.ORDERING,
FROM SYSIBM.SYSKEYS A,
SYSIBM.SYSINDEXES B
WHERE A.IXNAME = B.NAME
Page:163
AND
AND
AND
AND
AND
Page:164
A.IXCREATOR = B.CREATOR
A.IXCREATOR = B.TBCREATOR
B.TBNAME = <TAB NAME>
B.TBCREATOR = <TAB CREATOR>
B.UNIQUERULE = P
How do I get the list of affected programs due to a table change?

SELECT GRANTEE "PROGRAM", TTNAME "TABLE NAME", SELECTAUTH "SELECT",
INSERTAUTH "INSERT", UPDATEAUTH "UPDATE", DELETEAUTH "DELETE"
FROM SYSIBM.SYSTABAUTH
WHERE GRANTEETYPE= 'P'
AND TCREATOR= <TABLE CREATOR NAME>
AND TTNAME = <TABLE NAME>
ORDER BY PROGRAM
To get the affected Plan names if an index is dropped
SELECT BNAME, BCREATOR, BTYPE, DNAME
FROM SYSIBM.SYSPLANDEP
WHERE BTYPE = I
AND BNAME = <Index name to be Dropped>
SYSPLANDEP table records the dependencies of plans on tables, views, aliases, synonyms,
table spaces, indexes, functions, and stored procedures.
BNAME
- The name of an object the plan depends on.
BTYPE
- Type of object identified by BNAME:
A Alias, F User-defined function or cast function, I index, S Synonym
T table, P Partitioned table space, R-Table space, V- Views, O-Stored Procedure
DNAME
- Name of the plan.
For package dependency, query SYIBM.SYSPACKDEP.
To get the view definition
SELECT NAME, CREATOR, SEQNO, TEXT
FROM SYSIBM.SYSVIEWS
WHERE NAME = <VIEW NAME>
SEQNO - IF THE VIEW DEFINITION IS MORE THAN ONE LINE
TEXT - VIEW DEFINITION
To get the Table/VIEW/Alias details
SELECT NAME, TYPE, CREATOR, COLCOUNT, TBCREATOR, TBNAME,
KEYCOLUMNS, CREATEDTS, ALTEREDTS, CREATEDBY
FROM SYSIBM.SYSTABLES
WHERE NAME = <TABLE or View or Alias Name>
Page:165
AND CREATOR = < TABLE or View or Alias Creator>

TYPE = T (Table) or A (Alias) V (View) G (Temporary Table)
COLCOUNT - NO OF COLUMNS
- only for Tables AND Views (0 for Alias)
RECLENGTH - Record length
- only for Tables (0 for Alias and Views)
TBCREATOR (FOR ALIAS)
- ACTUAL TABLE CREATOR
TBNAME (FOR ALIAS)
- ACTUAL TABLE NAME
KEYCOLUMNS
- No of Primary Keys
How do I get all the objects used by a plan?

SELECT BNAME, BTYPE FROM SYSIBM.SYSPLANDEP WHERE DNAME=PLAN-NAME
How do I get all the indexes for the table?
SELECT NAME, COLCOUNT, UNIQUERULE, CLUSTERING FROM SYSIBM.SYSINDEXES
WHERE TBNAME=TABLE-NAME
UNIQUERULE P (Primary) U (Alternate)
CLUSTERING YES for clustering index.
Give an example for Self Join
Employee table has Employee No, Employee Name, Manager Name and other details
of employee. I want to know all the employees and their managers.
SELECT A.EMPNAME, B.MANAGERNAME
FROM EMPTABLE A, EMPTABLE B
WHERE A.MANAGERID = B.EMPNO
How do I get the difference between two timestamps in number of seconds?
SELECT (((DAYS (TS2) - DAYS (TS1)) * 86400) +
((HOUR (TIME (TS2)) - HOURS (TIME TS1))) * 3600) +
((MINUTE (TIME (TS2)) - MINUTE (TIME (TS1))) * 60) +
(SECOND (TIME (TS2)) - SECOND (TIME (TS1))))
FROM SYSIBM.SYSDUMMY1;
In the first part, you subtract the start time from the end time, deriving the number of
days between the timestamps. You then multiply the number of days by 86400, the
seconds in a day, to convert the days to the number of seconds.
SELECT (DAYS (TS2) - DAYS (TS1)) * 86400
In the second part, you derive the additional time (hours, minutes, and seconds, list in
hhmmss format) between the timestamps and convert the time to the number of
seconds. First you need to extract and convert the hours to seconds:
SELECT (HOUR (TIME (TS2)) - HOURS (TIME TS1))) * 3600
Next you need to extract and convert the minutes to seconds:
SELECT (MINUTE (TIME (TS2)) - MINUTE (TIME (TS1))) * 60
Then you extract the remaining seconds:
SELECT SECOND (TIME (TS2)) - SECOND (TIME (TS1))
Lastly, you add all partial results to produce the difference between the two
timestamps in total number of seconds.
Selected DB2-command panel commands: (Option 7 of DB2I)
Page:166
-DIS DB(DSNDB01)
Displays all the components in the Database DSNDB01 and their status.
-DIS DB(DSNDB01) SPACENUM(xxx)
Displays information about specific table space/index space(xxx) of database
-STA DB(DSNDB01)
Starts the database DSNDB01
-STA DB(DSNDB01) SPACENUM(xxx)
Start a specific tables pace/index space of a database.
SPUFI- SQL Processing Using File Input.

Option 1 of DB2I panel takes you to SPUFI panel.
__________________________________________________________________________________
|
|
| DSNESP01
SPUFI
SSID: DSN
|
| ===>
|
| Enter the input data set name:
(Can be sequential or partitioned)
|
|
1 DATA SET NAME..... ===> SMSXL86.DB2.SPUFI.INPUT(MEM1)
|
|
2 VOLUME SERIAL..... ===>
(Enter if not cataloged)
|
|
3 DATA SET PASSWORD. ===>
(Enter if password protected)
|
|
|
| Enter the output data set name:
(Must be a sequential data set)
|
|
4 DATA SET NAME..... ===> SMSXL86.DB2.SPUFI.OUTPUT
|
|
|
| Specify processing options:
|
|
5 CHANGE DEFAULTS... ===> Y
(Y/N - Display SPUFI defaults panel?)
|
|
6 EDIT INPUT........ ===> Y
(Y/N - Enter SQL statements?)
|
|
7 EXECUTE........... ===> Y
(Y/N - Execute SQL statements?)
|
|
8 AUTOCOMMIT........ ===> Y
(Y/N - Commit after successful run?)
|
|
9 BROWSE OUTPUT..... ===> Y
(Y/N - Browse output data set?)
|
|
|
| For remote SQL processing:
|
| 10 CONNECT LOCATION
===>
|
|
|
|
|
| PRESS: ENTER to process
END to exit
HELP for more information
|
|
|
|___________________________________________________________________________________
SMSXL86.DB2.SPUFI.INPUT(MEM1) It is my dataset where I am going to code

SQL statements I want to execute. If I already have my query in this dataset, then
I will set EDIT INPUT option as N.
SMSXL86.DB2.SPUFI.OUTPUT - Output of the SQL statement is received here. We
no need to allocate the dataset.
If AUTOCOMMIT is Y, then all the changes made to DB2 using SQL statements are
committed on successful execution of the query.
If you set the option CHANGE DEFAULTS to Y, then you will get the screen with
SPUFI default values. The SPUFI allows you to set the values for:
Page:167
1. Isolation level and Maximum number of lines to be returned from a SELECT

2. Output dataset characteristics.
3. Output format characteristics. (Defining length of numeric and character
column, defining headers in your output file)
QMF (Query Management Facility)
Most of the installations use SPUFI for pre-testing the SQL statements.
SPUFI comes with DB2 product. Some installation use QMF for their SQL operations.
QMF provides greater flexibility and lot of options in generating the result of the
query in a formatted output of report style.
Cobol-DB2 Compilation JCL

//APIJ1DB JOB (TVD2TM,3B42),COND=(4,LT),
//
MSGCLASS=Y,CLASS=X,NOTIFY=&SYSUID
//*********************************************************************
//* INSTREAM PROCEDURE FOR COMPILATION PRECOMPILE STEP
//*********************************************************************/
/PRECOMP EXEC PGM=DSNHPC,PARM='HOST(COBOL)',REGION=4096K
//DBRMLIB DD DISP=OLD,DSN=APIJ1.TSTINDIA.DBRMLIB(PSELECT2)
//STEPLIB DD DISP=SHR,DSN=D2TB.DSNEXIT
//
DD DISP=SHR,DSN=D2TB.DSNLOAD
//SYSIN DD DISP=SHR,DSN=APIJ1.TSTINDIA.JCL(PSELECT2)
//SYSCIN DD DISP=(NEW,CATLG,CATLG),
//
DSN=APIJ1.PRECOMP.LIST,UNIT=SYSDA,
//
SPACE=(800,(500,500))
//SYSLIB DD DISP=SHR,DSN=APIJ1.TSTINDIA.COBOL
//
DD DISP=SHR,DSN=APIJ1.TSTINDIA.DCLGEN
//SYSTERM DD SYSOUT=*
//SYSUT1 DD SPACE=(800,(500,500),,,ROUND),UNIT=SYSDA
//SYSUT2 DD SPACE=(800,(500,500),,,ROUND),UNIT=SYSDA
//*
//*********************************************************************/
/* INSTREAM PROCEDURE FOR COMPILATION COMPILE STEP
//*********************************************************************
//COBCOMP EXEC PGM=IGYCRCTL,COND=(4,LT,PRECOMP)
//SYSIN DD DSN=&&DSNHOUT,DISP=(OLD,DELETE)
//SYSLIN DD DSN=&&LOADSET,DISP=(MOD,PASS),UNIT=SYSDA,
//
SPACE=(800,(&WSPC,&WSPC))
//SYSUT1 DD SPACE=(800,(&WSPC,&WSPC),,,ROUND),UNIT=SYSDA
Page:168
//*********************************************************************
//* INSTREAM PROCEDURE FOR COMPILATION LINKEDIT STEP
//*********************************************************************
//LKED EXEC PGM=IEWL,PARM='XREF',
//
COND=((4,LT,COBCOMP),(4,LT,PRECOMP))
//SYSLIB DD DSN=SYS2.COB2LIB,DISP=SHR
//
DD DISP=SHR,DSN=D2TB.DSNLOAD
//SYSLIN DD DSN=&&LOADSET,DISP=(OLD,DELETE)
//SYSLMOD DD DISP=OLD,DSN=&USER..TSTINDIA.LOADS(&MEM)
//SYSUT1 DD SPACE=(1024,(50,50)),UNIT=SYSDA
//
PEND
//*
//*********************************************************************
//*INVOKING INSTREAM PROCEDURE FOR COMPILING COBOL-DB2 PROGRAM.

//*********************************************************************
//COMP EXEC DSNHCOB,
//
MEM=DYNSQL02,
//
USER=APIJ1,
//
PARM.PRECOMP='HOST(COBOL),QUOTE,SOURCE,XREF',
//
PARM.COBCOMP='QUOTE,SOURCE,NODYNAM',
//
PARM.LKED='LIST,XREF'
//PRECOMP.SYSIN
DD DSN=&USER..TSTINDIA.COBOL(&MEM),DISP=SHR
//PRECOMP.DBRMLIB DD DSN=&USER..TSTINDIA.DBRMLIB(&MEM),DISP=SHR
//PRECOMP.SYSLIB DD DSN=&USER..TSTINDIA.DCLGEN,DISP=SHR
//LKED.SYSLMOD
DD DSN=&USER..TSTINDIA.LOADS(&MEM),DISP=SHR
//LKED.SYSIN
DD *
INCLUDE SYSLIB(DSNELI)
//*********************************************************************
//*BIND-PLAN STEP.
//*********************************************************************
//BIND EXEC PGM=IKJEFT01,DYNAMNBR=20
//STEPLIB DD DSN=D2TB.DSNLOAD,DISP=SHR
//SYSOUT
DD SYSOUT=*
//DBRMLIB DD DSN=APIJ1.TSTINDIA.DBRMLIB(DYNSQL02),DISP=SHR
//SYSTSIN DD *
DSN SYSTEM(D2TB) RETRY(2)
BIND PLAN(APIJ1P1) MEMBER(DYNSQL02) OWNER(APIJ1) ISOLATION(CS) ACTION(REPLACE)
/*
Sample BIND Package Card and respective BIND Plan card WITH RUN command
//BIND
EXEC PGM=IKJEFT01,DYNAMNBR=20
//DBRMLIB DD DSN=DP.DBRMLIB,DISP=SHR
//SYSTSIN DD *
DSN SYSTEM(DSNP)
BIND PACKAGE(SOCCOLL) MEMBER(SOC02P01) OWNER(SOC) +
ACTION(REPLACE)
VALIDATE(BIND)
ISOLATION(CS)
END
Page:169
//*
//STEP1
EXEC PGM=IKJEFT01,DYNAMNBR=20,COND=(4,LT)
//SYSTSIN DD *
DSN SYSTEM(DSNP)
BIND PLAN(SOCBATCH)
PKLIST(SOCCOLL.*)
ACTION(REPLACE) +
RETAIN
ACQUIRE(USE)
RELEASE(COMMIT) +
ISOLATION(CS)
VALIDATE(BIND)
NODEFER(PREPARE) +
CACHESIZE(0)
END
//*
//JS01
EXEC PGM=IKJEFT01,DYNAMNBR=20,COND=(4,LT)
//SYSTSIN
DD *
DSN SYSTEM(DSNP)
RUN PROGRAM(SOC02P01) PLAN(SOCBATCH) LIB('DP.GOLIB')
END
//
Page:170
Customer Information Control System

CICS
Page:171
CICS (Customer Information Control System)

History
IBM launched the initial version of CICS in 1968.
It is a Database/Data Communication Control System where an application program
can concentrate on the application processing without worrying about OS, hardware
and others. Initially CICS was on macro level and later upgraded to command level.
This book is written based on CICS/MVS V2R3
Task and Transaction
Task is a unit of work and transaction is an entity that initiates the execution
of task. The transaction identifier identifies the transaction in CICS.
Conversation and Pseudo conversation
Program can communicate with user by a pair of SEND and RECEIVE
commands. But in between the SEND and RECEIVE, that is during the human
response time, all the resources are held by the program.
Once the user entered the information in the screen, the program proceeds
further. This mode of communication is called Conversational mode. It is un-famous
for the resource wastage.
In Pseudo conversation mode, whenever there is a need for conversation with
user, the program logically terminates there, releases the resources held by it and
pass control information to next transaction and the next transaction is automatically
started once user entered the information in the screen.
It is actually multi task operation but looks like conversation from user point
of view.
For easy understanding we would say the communication in telephone line is
conversational mode (Telephone line is in usage through out the communication.)
IRC as Pseudo conversation (Once the message sent, the line is freed and it again
gets the resources when the other side replied.)
Multitasking and Multithreading
Operating system allows execution of more than one task concurrently and
this is called multitasking. If the one or more concurrent tasks use the same copy of
the program, then this is called multithreading. So it is obvious that multitasking is a
subset of multithreading.
Reentrant and Quasi-reentrant
If the same copy of the program is used by multiple tasks, then the system
should take care of the proper reentrance after the SVC/CICS interruption.
Reentrant program is defined as a program that does not modify itself so that
it can reenter to itself and continue processing after an interruption by the SVC call
of OS. Batch programs are non-reentrant. Reentrancy under CICS environment is
called quasi entrant as the interruption in CICS may involve more than one SVC calls
or no SVC at all.
Page:172
COBOL programs should be compiled with RENT option for reentrancy in

multithreading environment.
CICS Control Program and Control Tables
CICS Nucleus consists of IBM-Supplied CICS control programs and
corresponding user-specified CICS control tables. The important tables with respect
to developer point of view are given below:
Control
Table
PCT
PPT
FCT
DCT
TST
RCT
SNT
TCT
PLT
JCT
Function
The transactions and the main program associates with the transaction
should be registered in Program table. Task control Program (KCP)
refers PCT.
All the CICS programs and Maps have to be registered in Processing
Program Table. Program Control Program (PCP) refers PPT.
All the VSAM files used in the CICS programs has to be registered in
File Control Table. File Control Program (FCP) refers FCT.
Transient data queues should be predefined in Destination Control
Table. Transient Data Program refers DCT.
If you want to recover Temporary storage queues during system crash,
then they should be registered in Temporary Storage Table.
If any DB2 commands are used in the program, then the PLAN should
be registered here.
User ID and Password should be registered in Sign-On-Table.
All the terminals should be registered in Terminal Control Table.
All the programs that need to be automatically started during CICS
start up and Shut down should be listed in Program List Table.
Control information of system logs and journal files is stored in Journal
Control Table. Journal Control Program refers to JCT.
CICS-DB2-COBOL Program Compilation

COBOL programs cannot recognize CICS commands. So all the CICS
commands are coded within EXEC CICS and END-EXEC scope. The program is first
passed to CICS Translator and the translator will convert all the CICS commands to
COBOL call statements and this modified source is passed to COBOL compiler.
If your Program has DB2 commands then the sequence of preparing load
module would be DB2 Pre-compiler, CICS Translator, COBOL Compiler and Link editor.
Logically speaking the order of DB2 precompiler and CICS translator can be reversed
also. But as a convention we are first doing precompilation.
If translation is done first, CICS translator tries to recognize to DB2
statements and would issue diagnostic messages. The other reason is, most of the
programs do lot of DB2 operations than CICS operations. So if you do precompilation first, all the DB2 statements are converted into COBOL call statements in
first phase itself and the translation time would be less.
COBOL-CICS programs should be compiled with RENT RESIDENT NODYNAM
and LIB options.
General Syntax of CICS statement
Page:173
EXEC CICS function

[(option (argument value)]
[(option (argument value)]
END-EXEC.
Restricted COBOL commands in CICS environment
OS SVC Triggering statements
ACCEPT CURRENT-DATE DATE DAY DISPLAY
I/O statements
OPEN CLOSE READ WRITE REWRITE DELETE

START
RETURN RELEASE
SORT statement
VS COBOL 2 allows STOP RUN and that returns control back to CICS.
MAP DESIGN
Before get into program design, let us see how maps (screens) are designed
in CICS. Most of the installations use tools like SDF for screen designing. The tools
generate BMS macros for the designed screen. We brief the BMS macros involved in
the map design. BMS is acronym for Basic Mapping Support.
MAP and MAPSET
A screen designed thru BMS is called MAP. One or a group of maps makes up
a MAPSET.
PHYSICAL MAP AND SYMBOLIC MAP
Physical maps control the screen alignment, sending and receiving of
CONSTANTS and data to and from a terminal. They are coded using BMS macros,
assembled and link edited into CICS LOAD LIBRARY. They ensure the device
independence in application programs.
Symbolic maps define the map fields used to store the VARIABLE data
referenced in COBOL program. They are also coded used BMS Macros. But after
assembling, they are placed in a COPY library and then copied into CICS programs
using COPY statement. They ensure device and format independence to the
application programs.
BMS Macro Coding Sheet
Since BMS map definitions are purely assembler macros, the following coding
convention must me maintained.
1
10
16
(1-8) Label (10-15) Macro name
(16-71) Operands
72
72-Continuation.
Macro name. There are three macros used in BMS coding.

1. DFHMSD. Defining Mapset
2. DHMDI. Defining Maps
3. DFHMDF. Defining Fields.
All the BMS coding starts with DFHMSD followed by one or more DFHMDI.
One or more DFHMDF follow DFHMDI.
Page:174
Label: Label is name of the operation. If the macro is DFHMSD, it specifies

map-set name. If the macro is DFHMDF then it specifies field name.
Operands: Define the parameters for the macro being invoked.
Continuation: Continuation of macro is achieved by X in 72nd column of the
current line and continues from the 16th column of next line.
Comments: Comment lines are indicated by * in column 1. A comment line
cannot be placed between continuation lines and comment lines cannot be continued.
Blank lines: You cannot have blank lines in the assembler program.
DFHMSD
It is used to define a mapset with its characteristic and to end a mapset. So
you will find two DFHMSD in any BMS coding. The important operands are below:
TYPE.
It should be DSECT for Symbolic map generation, MAP for physical map
generation and FINAL to indicate the end of mapset. Alternatively symbolic
parameter &SYSPARM can be coded in the TYPE of defining DFHMSD and the value
can be overridden in the PARM parameter of assembly procedure. This avoids the
change in the BMS coding.
MODE.
IN for input maps like order entry screens and OUT for output maps like
display screens and INOUT for input-output maps like update screens.
LANG.
It specifies the language in which the symbolic map is to be generated. It can

be COBOL, PLI, ASM or RPG.
STORAGE.
AUTO is used to acquire separate symbolic map area for each mapset.
BASE=MAP-IOAREA allows multiple maps from more than one mapset to share
same storage area. MAP-IOAREA will be redefined multiple times to achieve it.
TIOAPFX.
It should be YES to reserve the prefix space of 12 bytes for BMS commands
to access TIOA properly. This is required for command level CICS.
CTRL.
Device control requests are placed here. Multiple parameters are separated by
comma. FREEKB is used to unlock the keyboard. FRSET is used to reset the MDT of
all the fields in all the maps to zero. ALARM is used to set an alarm at screen display
time. PRINT is used to send the mapset to printer.
TERM.
If anything other than 3270 terminal is used for display of screens, then it
should be coded here. This ensures device independence by means of providing the
suffix. SUFFIX is used to specify suffix for the terminal and it should correspond to
TCT entry of the terminal.
DFHMDI
Page:175
It is used to define a map with its characteristic in a mapset. There can be

any number of DFHMDI. Some of the important operands of DFHMDI are below:
SIZE.
It has two arguments namely length and breadth and as a whole the
size of the map is specified here.
LINE.
The map starting line is mentioned here.
COLUMN.
The map-starting column within the LINE is mentioned here.
JUST.
RIGHT or LEFT is coded here to inform the justification of the map
within mapset.
The above four parameters decides the size and location of map within map
set. CTRL and TIOAPFX can be also coded in DFHMDI. Value of DFHMDI overrides the
value of DFHMSD.
DFHMDF
It is used to define a field with its characteristic in a map. There can be any
number of DFHMDF within DFHMDI. Some of the important operands of DFHMDF are
below:
POS.
It has two arguments that decided the position of the field. The two
arguments are line and column. It is the position where the attribute byte of the field
starts.
LENGTH.
The length of the field is coded here. It excludes the attribute character.
ATTRIB.
All the input and output fields are prefixed by one byte attribute field that defines
the attributes of the field. Some of the attributes are:
1. ASKIP/PROT/UNPROT
Mutually exclusive parameters that define the
type of the field. UNPROT is coded for input and input-output fields. PROT is
coded for output and stopper fields. ASKIP is coded for screen literals and
skipper fields. The cursor automatically skipped to next field and so you
cannot enter data into skipper field.
2. NUM.
0-9,Period and are the only allowed characters.
3. BRT/NORM/DRK
Mutually exclusive parameters that define the
intensity of the field.
4. IC
Insert Cursor. Cursor will be positioned on display
of map. If IC is specified in more than one field of a map, the cursor will be
placed in the last field.
5. FSET
Independent of whether the field is modified or
not, it will be passed to the program. MDT is set for the field.
JUSTIFY.
RIGHT is the default value. Code LEFT for numeric fields.
PICIN and PICOUT.
It defines the Picture clause of the symbolic map in COBOL and useful for
numeric field editing.
INITIAL.
Page:176
The default value of the field is coded here. When the MAP is sent, this value
will appear in the field. The constant information like TITLE is coded using INITIAL
keyword of field definition. To avoid data traffic, these constant information fields
should not be coded without LABEL parameter. If there is no LABEL parameter, then
symbolic map will not generated for those fields as they are unnamed fields.
How a field looks like in a Symbolic map

Let the label of one DFHMDF is EMPNAME in the map EMPDET and the length
of it is 20. The respective symbolic map would look like follows.
01 EMPDETI.
02 FILLER PIC X(12)
02 EMPNAMEL PIC S9(04) COMP.
02 EMPNAMEF PIC X.
02 FILLER REDEFINES EMPNAMEF.
03 EMPNAMEA PIC X.
02 EMPNAMEI PIC X(20).
Other fields.
01 EMPDETOO REDEFINES EMPDETI.
02 FILLER PIC X(12)
02 FILLER PIC X(03)
02 EMPNAMEO PIC X(20).
Other fields
TIOAPFX = YES creates this 12 byte filler.

Length Field
Flag byte
Attribute byte
Actual field (Input)
TIOAPFX = YES creates this 12 byte filler.
Actual field (Output)
So four fields are generated for every named field in the BMS.
Fieldname+L
It has length of the field entered by the user during the input operation.
Fieldname+I
It is the actual input field that carries the entered information. The value of
this field is X00 if no data is entered. The space corresponds to X40.
Fieldname+A
It is attribute byte. It defines the attributes of the field.
Fieldname+F
It is a flag byte. It has X00 by default. It will be set to X80 if the user
modifies the field but no data is sent. That is when the user pressed clear key over
the field.
Fieldname+O
This field should be populated in the program before sending the screen to
display.
Please note that the words INPUT (RECEIVE) and OUTPUT (SEND) are with respect to
program.
Page:177
More on Attribute field

We have just seen that attribute field is of 1 byte that specifies the attributes
of the field. One byte is of eight bits and the values in each bit has its own meaning.
For example,
Bit 2 - 0 indicates the field is unprotected and 1 indicates the field is protected.
Bit 3 0 indicates the field is alphanumeric and 1 indicates the field is numeric.
Bit 2 and 3 11 indicates that the field is Auto-skip.
Bit 4 and 5 00 indicates Normal intensity and non-detectable.
01 indicates Normal intensity and detectable.
10 indicates High intensity and detectable.
11 indicates Dark and Non-detectable.
Bit 7 0 indicates MDT is OFF 1 indicates MDT is ON.
Modified Data Tag(MDT)

MDT is one bit field of the attribute byte. The program can receive only the
fields with MDT 1 on RECEIVE. Effective use of MDT can reduce the data traffic
drastically in the communication line.
MDT can be SET/RESET in the following ways.
1. When the user modifies the field, the MDT of the field is automatically set
to ON.
2. CTRL=FRSET of DFHMSD or DFHMDI will RESET the MDT to OFF for all
the fields in the mapset or map. FSET keyword of the attribute operand
definition of DFHMDF will set the MDT to ON . It overrides the FRSET
definition for the specific field.
3. Before sending the screen, by overriding the MDT bit of attribute byte of
field the MDT can be set to ON.
If you are specific on the values of some fields independent of whether the
user has modified or not, code them with FSET. One good example is default values
for the input fields (like Order received date). If the user finds default value (current
date) in the screen and it is fine with his requirement, then he wont touch the field
and MDT will not be set. The program cannot receive the field as MDT is OFF.
But actually the program needs this field. So this field should have been defined with
FSET.
CURSOR Positioning
Positioning of cursor is an important area in screen design. By default, the
cursor will be placed in the first unprotected field.
Static Positioning.
If IC is coded in the attribute operand of DFHMDF macro, then the cursor will
be placed in that field. If IC is coded in more than one field, then last field with IC
will get the cursor.
Symbolic Positioning.
Page:178
Move 1 to length of the field where you want place the cursor and send the
map with CURSOR option. This is device independent method and recommended to
control the cursor position dynamically.
Relative Positioning.
Send the map with CURSOR(value) option. This will set the cursor at the
position coded by value, relative to the first column of the screen. This is device
dependant method of dynamic cursor positioning and not recommended. When there
is change in screen layout, the program needs to be modified.
CURSOR(30) will place the cursor in the 30th column. (First line first column is
ZERO).
Cursor Position.
EIBPOSN of DFHEIBLK contains the offset of cursor position in the screen
when the data was transferred to program from the terminal.(relative to zero). It is
half word binary field.
Dynamic attribute setting

Any attribute of the field can be modified in the program by setting the bits of
attribute field properly. Changing the attribute in the program in a dynamic method.
CICS provides the standard attribute character list in the form of copybook.
(DFHBMSCA). This can be copied into our application program using COPY statement
and we can use the variables in the copybook to change the attribute of a field in the
program.
For example, if the information entered by the user in a particular field is
failed in validation, then as a usual practice, in addition to throw an error message in
the bottom of the screen, you may wish to hi-light the field that is in error.
MOVE DFHUNIMD to FIELDA. == > Will achieve this.
SKIPPER and STOPPER field
Skipper is unlabelled one byte field with the auto-skip attribute and stopper is
unlabelled one byte field with protected attribute.
Skipper field skips the cursor to next field when the current field is filled up to
its full length, whereas stopper field stops entering any more character after the
length of the field is reached. The user has to press RESET key to proceed further.
Skipper field is used to speed up the user entry and recommended to code at
the end of each unprotected field. Stopper field is coded based on requirement and
importance of the field. Stopper field usually follows the last unprotected field in the
screen.
BMS INPUT-OUTPUT OPERATIONS
Basic BMS input-output operations are carried out using the following
commands. RECEIVE MAP, SEND MAP, SEND CONTROL, SEND TEXT, SEND PAGE.
RECEIVE MAP.
It is used to receive the information entered by the user into program. If
INTO is not coded, then CICS automatically finds the symbolic map area
(mapname+I) and places the mapped data. The values of the field can be read in the
program by referring to fieldname+I.
Page:179
EXEC CICS RECEIVE MAP(map-name) MAPSET(mapset-name) END-EXEC.

The option TERMINAL ASIS overrides the upper-case translation specified in
the TCT. As we already discussed, if the user didnt modify any of the fields then
MDT will be OFF for all the fields. Zero bytes transmission results MAPFAIL exception
error.
SEND MAP.
This is used to send a map to a terminal. Before issuing this command, the
application program should prepare the data in the symbolic map area. If FROM is
not coded, then CICS automatically finds the symbolic map area (mapname+O) and
takes the data to the terminal.
EXEC CICS SEND MAP(map-name) MAPSET(mapset-name) END-EXEC.
Other Options in SEND MAP are:
ERASE. Erases the screen before displaying the MAP. When you first time
throw the screen, it should be specified with ERASE. It will not be coded when you
just want to display the error messages over the current screen. ERASEUP erases
only the unprotected fields of the screen before displaying the MAP.
DATAONLY. Only symbolic map data is sent to the terminal.
MAPONLY. Only physical map data is sent to the terminal. FROM cannot be
coded.
FREEKB, ALRAM and FRSET can be also coded and the meaning is already discussed.
TEXT Building- SEND-TEXT, ACCUM and PAGING
Text streams can be sent to the terminal without any pre-defined BMS maps.
This is called text building. Text streams can optionally have header and trailer.
Syntax the SEND TEXT command is:
EXEC CICS SEND TEXT
FROM(data-value)
LENGTH(data-value)
[HEADER(data-value)]
[TRAILER(data-value)]
[ERASE] [ACCUM] [PAGING]
END-EXEC.
When to code ACCUM?
The text stream can consist of multiple paragraphs and the building of
message can happen in multiple stages. If you want to send the text stream only
after the building of all the paragraphs, use ACCUM option for the SEND-TEXT of
every individual paragraph being built. They are accumulated and will not be sent
immediately.
Once the page is built with complete message of all the paragraphs, issue
SEND PAGE command that will send the complete text-stream to the terminal.
HEADER is placed in the first SEND-TEXT and TRAILER is placed in the SEND PAGE.
When to code PAGING?
ACCUM alone is enough if the text-stream can fit within a single page.
But if the text stream is expected to exceed one page, then code PAGING in all the
SEND-TEXT commands in addition to ACCUM option. The last SEND-PAGE command
sends the first page of message to the terminal.
The user can enter paging commands or keys to walk through the pages.
Page:180
To use keys instead paging commands, the keys should be mapped with paging
commands in the System Initialization Table (SIT). Usually there will be a mapping
for PF7 and PF8 with page up and page down commands respectively. PURGE
MESSAGE is used to purge the message that is accumulated but not yet send.
Format of HEADER and TRAILER
If TEXT-HEADER is the variable used in HEADER command and the header is
ORDER INVENTORY Page nn- , then the variable TEXT-HEADER should be defined
as follows in working-storage section.
01 TEXT-HEADER.
05 FILLER PIC S9(4) COMP VALUE 27. => Length of the text.
05 FILLER PIC X VALUE &.
=> Character identifying automatic
page number in the text.
05 FILLER PIC X.
=> One byte control field used by BMS.
05 FILLER PIC X(27) VALUE ORDER INVENTORY Page && - .

=> Actual text in the header.
TRAILER also can be coded in the same way. Automatic- page-numberidentifying character will be spaces for TRAILER working storage variable.
ACCUM and PAGING can be also used with SEND-MAP and the meaning is
same.
SEND-CONTROL
When the MAP is sent, we can specify device control options like FREEKB
ALARM ERASE, along with SEND-MAP. If one wants to issue the device control
options before sending the map, SEND CONTROL will be useful. It is used to
establish the device control options dynamically.
Syntax:
EXEC CICS SEND CONTROL
[CURSOR(data-value)] [ERASE | ERASEUP] [FREEKB] [ALARM] [FRSET]
END-EXEC
PRINT Through BMS
PRINT option of SEND MAP command allows the reports to be printed at the
local printer connected to the terminal. In this case, MAP is sent to the printer and
not to the terminal.
Syntax:
EXEC CICS SEND MAP(MAP-NAME)
MAPSET(MAPSET-NAME)
PRINT
NLEOM
END-EXEC.
NLEOM option is recommended with PRINT option. When this option is used,
1. BMS builds the data stream using new line characters and blank characters to
position the fields on the printer page.
2. The data stream is terminated by EOM (end of message) that stops printing.
So printer buffer is efficiently used, allows larger pages to be printed from the
same buffer that makes the printing faster.
Page:181
Message Routing
A message can be routed to one or more terminals other than the direct
terminal with which the program has been communicating. The message eligible for
message routing is, a message constructed by the SEND MAP command with the
ACCUM option.
ROUTE command establishes the message routing environment and the SEND
PAGE command issued after ROUTE command sends the message to the destination.
Syntax:
EXEC CICS ROUTE [LIST(data-area)], [OPCLASS(data-area)],
[INTERVAL(hhmmss)|TIME(hhmmss)],
[TITLE(data-area)], [ERRTERM(name)]
END-EXEC.
LIST and OPCLASS name the route list and operator class codes respectively.
INTERVAL / TIME determines the actual timing of message delivery in the time
interval or the time respectively.
TITLE names the title field defined in the working storage section and
ERRTERM specify the terminal ID where the error message (if any) to be sent.
Route list is prepared in working storage using the following convention.
TTTTrrOOOsrrrrrr 8 bytes named r are declared as spaces. TTTT names the
terminal identifier as in TCT and OOO specify the operator id as in SNT. s is status
flag. Code as many 16 bytes fields as the destinations and indicate end of route list is
with the declaration of half word binary field with 1 value.
The message can be routed to every terminal at which users of the specified
operator class is signed on. This is done using OPCLASS.
Program Design
Pseudo logic in CICS programs is achieved in three ways.
1.Multiple Programs and Multiple Transaction Ids.
The conversation program is logically and physically divided into multiple
programs and each program is registered with one transaction ID. Each program
issues RETURN command with next transaction ID. So after the screen is sent, the
first program will be terminated. Once the user filled up the information in screen
and press ENTER, the next program of the next transaction ID (returned by the first
program) gets control and the same process is followed for n programs.
Advantage:
Easy development and maintenance.
Disadvantage: Possible Code redundancy and number of PCT and PPT entries.
2.Multiple Transactions and single program.
First paragraph of the procedure division controls the flow the different
sections of the program based on the transaction identifier stored in EIBTRNID.
Every section ends with RETURN command with next transaction to be invoked on
completion of the screen information by the user.
Thus the conversational program is logically divided than physical division.
Advantage:
PPT entries are less compared to case 1.
Disadvantage: Number of PCT entries are still the same as case 1.
3.One transaction and One program. (Preferred way)
Page:182
RETURN command has an optional parameter of COMMAREA and LENGTH.

Using this parameter we can pass information between the transactions.
So first paragraph of your program will check whether this is the first entry or
nth entry. If EIBCALEN is zero, this is first entry. So divert the program to the section
that will do all initial processing and throw the first screen for the user.
In addition to this it updates COMMAREA with some control information like
internal transaction identifier. This COMMAREA is passed along with RETURN
TRANSID. So EIBCALEN is updated with LENGTH of COMMAREA. The second and nth
entrances dont have EIBCALEN as zero and so based on the information in the
COMMAREA, transfer the control to different sections of the program.
Task Initiation Process
1.User entered transaction identifier.
2.Terminal Control Program along with Terminal Control Table recognizes the
incoming data from the terminal.
3.Storage Control Program acquires the storage for the terminal input output
area (TIOA)
4.Terminal Control Program places the data from the terminal into TIOA and
sets the pointer into TCT entry of the terminal.
5.If there is no task is associated with the terminal, then TCP gives the
control to task Control Program, which realizes the transaction identifier in the data
in TIOA. (KCP)
6.SCP acquires the storage area for task control area (TCA) in which KCP
prepares the control data for this task.
7.KCP refer PCT to identify the program associated with the transaction.
8.Program Control Program (PCP) loads the program into main memory from
the load library and gives the control back to KCP.
9.KCP gives the control to application program loaded into main memory.
10.Program started processing Task is initiated.
Ways of Data Passing between transactions.
1. By COMMAREA.
In the example below, Working Storage item WS-ITEM of length WS-LENGTH
is passed to the transaction EMPC. The program for the transaction EMPC should
have DFHCOMMAREA of size WS-LENGTH in the linkage section to receive the
information passed by RETURN of this program.
EXEC CICS
RETURN TRANSID(EMPC) COMMAREA(WS-ITEM) LENGTH(WS-LENGTH)
END-EXEC
WS-LENGTH is full word binary and the maximum length that can be specified
is 64K. So we can pass data of maximum size 64K. But it is recommended not to
pass more than 24K.
LINK and XCTL can also have use COMMAREA to pass the information to the
program being called and the concept is same.
2. Queues.
There are two types of queues TSQ and TDQ. TSQ can be used as scratch pad
in main memory. You can write as much as you want in TSQ and they will be
available to all the transactions that are aware of the queue name. TSQ is primarily
Page:183
used to share huge amount of information across the transactions. Refer Queues
section for more details.
3. TCTCA, TWA, CWA.
TWA Transaction work area One per task - Work area associated with the
task.
TCTUA Terminal Control Table User Area Work area associated with a
terminal and defined as one per terminal in TCT.
CWA - Common work area System work area defined by system
programmer in SIT. (System Initialization Table)
We can use TWA, TCTUA and CWA for sharing the information across the
transactions.
External address ability using BLL CELLS of OS/VS COBOL
Base Locator for Linkage is used to access the memory outside you working
storage. If it is used in input commands like READ, RECEIVE, the performance will be
good as we would be directly accessing the input buffer than working storage item.
If you want to access system areas like CWA, TCTUA or CSA that exist outside your
program, then you should use BLL.
Code Linkage section as below:
01 PARM-LIST.
05 FILLER
PIC S9(08) COMP.
05 TCTUA-PTR
PIC S9(08) COMP.
05 TWA-PTR
PIC S9(08) COMP.
01 TCTUA-AREA.
05 TCTUA-INFORMATION PIC X(m).
01 TWA-AREA.
05 TWA-INFORMATION PIC X(n).
The mapping between the pointers and data area is done in linkage section as
follows.
First filler is points the current 01 level, which is PARM-LIST itself.
TCTUA-PTR points to next 01 level, which is TCTCA-PTR
TWS-PTR points to next 01 level, which is TWA-PTR.
In the procedure division, code the following:
EXEC CICS ADDRESS TWA(TWA-PTR) TCTUA(TCTUA-PTR) END-EXEC.
SERVICE RELOAD TCTUA-AREA
SERVICE RELOAD TWA-AREA
This will store the pointer of TWA in TWA-PTR and TCTUA in TCTUA-PTR.
As the mapping between BLL cells and areas is already done in linkage section,
TCTUA-AREA can access m bytes of TCTUA, which exist outside your working storage
and similarly TWA-AREA to access n bytes of TWA, which exist outside your working
storage area.
To ensure the addressability, SERVICE RELOAD statement should follow
whenever the content of BLL cell is changed in anyway. So there should be a
SERVICE RELOAD statement after the ADDRESS statement as follows.
Page:184
Enhancement by VS COBOL2
ADDRESS OF operator of VS COBOL2 easies the access of external items.
The above requirement can be met in VS COBOL2 as follows. There is no need for
PARM-LIST or SERVICE RELOAD.
LINKAGE SECTION.
01 TCTUA-AREA.
05 TCTUA-INFORMATION PIC X(m).
01 TWA-AREA.
05 TWA-INFORMATION PIC X(n).
PROCEDURE DIVISION.
EXEC CICS ADDRESS TCTUA (ADDRESS OF TCTUA-AREA)
TWA(ADDRESS OF TWA-AREA)
END-EXEC.
ADDRESS OF maps the TCTUA-AREA with TCTUA exists outside your working
storage.
ASSIGN statement
It is used to access the system value outside of application program.
Some of them are length of common areas, user-id.
CWALENG, TCTUALENG, TWALENG assign to full word binary working storage item.
USERID
- assign to X(08) field.
EXEC CICS ASSIGN
USERID(WS-USERID)
END-EXEC
The above command fetches the user id into WS-USERID.
EIB-Executive Interface Block

When CICS translator translates your program, it adds DFHEIBLK copybook as
the first entry in your linkage section. Whenever a task is initiated, the task related
information could be accessed using the fields in the copybook. EIB is acronym for
Executive Interface Block.
Some of the fields in EIB are:
EIBCALEN
Length of DFHCOMMAREA.
EIBDATE and EIBTIME
Date and Time of task initiation.
EIBAID
- The last attention identifier pressed by the user.
EIBFN
- Function code of the last command. (2 bytes)
EIBRCODE
- Response code of the last command. (6 bytes)
EIBTRMID
- Terminal ID
EIBTRNID
- Transaction ID
CICS PROGRAM LINKAGE SECTION
If you look into the translated source listing, the linkage section of CICS
program might have the following.
1.
2.
3.
4.
Page:185
DFHEIBLK (Introduced by Translator)

DFHCOMMAREA (Coded by you as the first entry in linkage section.)
PARM-LIST (BLL cells coded by you)
Multiple 01 levels (The number of 01 entries is equal to number of BLL
cells you have coded in PARM-LIST excluding the first BLL cell that points
to PARM-LIST itself.)
GETMAIN and FREEMAIN.

GETMAIN command is used to obtain a certain amount of storage in the main
memory.
EXEC CICS GETMAIN SET(ADDRESS OF LK-ITEM)
LENGTH(data-value)|FLENGTH(data-value)
INITIMG(data-value)
END-EXEC.
LK-ITEM is a linkage 01 item whose length is equal to length mentioned in the
LENGTH parameter and that much memory will be allocated for you in the main
memory. The initial value of this field will be set by INITIMG parameter.
EXEC CICS FREEMAIN DATA(LK-ITEM)
END-EXEC
The above command will release the main memory occupied by you using GETMAIN.
These commands are highly useful in macro level coding where the quasireentrancy is the responsibility of the application programmer. Quasi-reentrancy is
guaranteed in the CICS command level COBOL programs and so the need and usage
of these commands are almost obsolete.
LINK XCTL and CALL

It is good practice to develop several functional modules rather than building
one lengthy program that does all the functions. It makes the program tough to
maintain, understand and debug. The size of the program will be large and it also
occupies more resources. Functional splitting of sub modules and calling or
transferring the control to them is achieved by LINK, XCTL and CALL.
CICS programs are usually run at various levels. The first program, which
gets the control from CICS, that is the program registered against the transaction in
PCT, runs at first level. Linked and Called programs are running one level lower than
linking or calling program and so the RETURN or GO BACK in these programs will
give the control back to Linking or calling program. XCTL programs run at the same
level and so the RETURN gives control to next upper level.
Syntax: EXEC CICS LINK|XCTL PROGRAM(name) COMMAREA(ws-area)
LENGTH(ws-length) END-EXEC.
Page:186
Ws-length is length of ws-area and ws-area has the information that needs to
be passed to LINK/XCTL program. VS COBOL 2 allows reentrant program and so
whatever is achieved using LINK or XCTL can be also achieved by COBOL CALL
statement. The called programs should be attached to the main program during link
edit and so the size of the load module will be large in this case.
Prefer XCTL whenever possible as it involves fewer overheads. If that is not
possible then based on sub-program size and possible modification, go for LINK or
CALL. If the size of the program is less and used by only very few programs go for
CALL as it wont increase the size much at the same time application will run fast.
If the program is expecting more changes, then calling program also need to
be compiled for every change in sub program. In case of LINK, the compilation of
linked programs is enough. Called CICS programs need not be registered in PPT
whereas LINK and XCTL programs should be.
LOAD and RELEASE
LOAD is used to load the program or table that is compiled or assembled, link
edited and registered in PPT. It is mostly used for loading the assembler table.
EXEC CICS LOAD PROGRAM(PGM1) SET (ADDRESS OF LK-ITEM)
LENGTH(WS-LENGTH)
END-EXEC
Program PGM1 is loaded and the address of the program or table is mapped
to LK-ITEM and so the table can be accessed using the linkage are LK-ITEM. The size
of the linkage item is WS-LENGTH. If HOLD keyword is coded in the LOAD
command, then loaded program or table will be permanently resident until explicitly
released. If it is not mentioned, then termination of the task release the program or
table.
EXEC CICS RELEASE PROGRAM(PGM1) END-EXEC is used to release the
program or table.
AID and HANDLE AID

There are certain attention identifiers (AID) associates with every screen.
Entry screens have ENTER and PF1 (Help) keys. Query screens have PF7 (Page Up),
PF8 (Page down), PF1 (Help) and so on. Based on the key pressed by the user, the
processing should happen in the program. EIBAID of DFHEIBLK has the last user
pressed key. It can be verified in the program after the RECEIVE command.
For verification of values in EIBAID, Copy the standard AID list copybook
provided by CICS in your program (COPY DFHAID). Now you can easily verify the key
pressed as follows.
EVALUATE EIBAID
WHEN DFHENTER PERFORM PARA-1
WHEN DFHPF1 PERFORM PARA-2
WHEN OTHER PERFORM PARA-3
END-EVALUATE.
Page:187
The above checking is done in COBOL and this kind of check has to be done
after every RECEIVE command to properly route the flow. CICS provides its own
routing processing based on the key pressed by HANDLE AID command. This will be
effective through out the program and reduce the code redundancy. The routing will
be automatically invoked on every RECEIVE.
EXEC CICS HANDLE AID
DFHENTER(PARA-1)
DFHPF1(PARA-2)
ANYKEY(PARA-3)
END-EXEC.
HANDLE CONDITION and NO HANDLE
Every CICS command has its own exception list. One example is MAPFAIL
exception of RECEIVE command. HANDLE CONDITION is used to transfer control to
the proper procedure on expected exceptions. HANDLE condition cannot track
program interrupted ABENDS like S0C4 or S0C7. It deals only with exceptions of
CICS commands.
The conditions handled are effective from where it appears to the end of the
program. One Handle Condition can be overridden by another condition. The main
program conditions will not be effective in sub-programs.
EXEC CICS HANDLE CONDITION
MAPFAIL(PARA-1)
PGMIDERR(PARA-2)
LENGERR(PARA-3)
ERROR(PARA-X)
END-EXEC.
= >Any error condition other than MAPFAIL,PGMIDERR and LENGERR will transfer
the control to PARA-X.
If you want to reset the diversion done by previous HANDLE condition, code
the condition name but dont mention the paragraph name. Maximum 12 conditions
can be coded in one HANDLE CONDITION statement.
If you want to deactivate all the conditions for a particular CICS command,
code NOHANDLE. There are possibilities for infinite loop when the CICS command in
the exception routine of the HANDLE CONDITION ends with same exception. In the
above example, if the there is LENGERR in PARA-3 for any of the CICS command
coded there, then the control again come to PARA-3 and forms infinite loop. In such
cases, NOHANDLE will be useful (in the ABEND routine CICS commands).
IGNORE CONDITION.
The syntax is same as HANDLE CONDITION but it causes no action to be
taken if the specified condition occurs in the program. The control will be returned to
the next instruction following the command, which encountered the exceptional
condition. It will be effective from the place where it appears to the end of the
program or until any HANDLE condition overrides it.
Maximum 12 conditions can be coded in one IGNORE CONDITION statement.
RESP
Like the file status verification of batch COBOL program, SQLCODE verification
of DB2 program, the success or failure of CICS command can be done using COBOL
statements and this give more structured look for the program.
Page:188
To achieve this, code RESP along with CICS command. CICS will place the
result into the variable coded in RESP. It can be checked against DFHRESP(NORMAL)
or DFHRESP(condition) and appropriately control the flow following the command.
EXEC CICS RECEIVE MAP(A) MAPSET(B) RESP(WS-RCODE) END-EXEC
EVALUATE WS-RCODE
WHEN DFHRESP(NORMAL) PERFORM PROCESS-PARA
WHEN DFHRESP(MAPFAIL) PERFORM PARA-1
WHEN OTHER PERFORM PARA-X
END-EVALUATE.
WS-RCODE should be defined as full word binary in working storage. If you
code RESP, then HANDLE CONDTION will not be effective for those CICS commands.
PUSH and POP.
They are used to suspend and reactivate respectively, all the HANDLE
CONDITION requests currently in effect. In real life programming, you may want to
deactivate all the active handle conditions and diversions for a subroutine (section)
embedded in the program. This can be achieved using PUSH HANDLE. When you
come out of that section, you can reactivate all the pushed commands and that can
be done using POP HANDLE.
HANDLE ABEND
HANDLE CONDITION command intercepts only abnormal conditions of the
CICS command execution whereas HANDLE ABEND takes care of any abnormal
termination within the program.
EXEC CICS HANDLE ABEND
LABEL(ABEND-ROUTINE)|RESET|CANCEL
END-EXEC.
LABEL is to activate an exit. CANCEL is used to cancel the previously
established HANDLE ABEND request. RESET is to reactivate the previously cancelled
HANDLE ABEND request.
USER-ABEND
In batch COBOL, the user ABENDS can be thrown by calling the assembler
routine ILBOABN0 using AB-CODE whereas AB-CODE is working storage field of half
word binary.
The following command is used to throw user ABENDS in CICS.
EXEC CICS ABEND ABCODE(9999)
END-EXEC is used to throw user ABEND 9999.
FILE-HANDLING.
CICS supports only VSAM and BDAM files. All the files, that you want to use in
your CICS application, should be registered in FCT with their complete attributes.
CICS commands refer the FCT entry name for doing operation on the file.
As all the files are already declared and defined in tables, coding File-Control
Paragraph of ENVIRONMENT division or FILE SECTION of DATA DIVISION is
meaningless and not required. Thus CICS frees the application program from any
data dependant coding.
Page:189
The unit of IO during READ is one control interval. So even you read one
record into your program, the complete control interval is read into main memory
buffer. The size of control interval is preferred to be large for sequential processing
and small for random processing.
The file should be OPEN to issue an I-O command. The status of the file can
be queried using the master transaction CEMT. It can be OPENED, CLOSED, ENABLED
or DISABLED. This explicit opening or closing can be done using CEMT.
But this needs human intervention.
CICS Version 1.7 introduces automatic opening of the file if it is not in open
status during the access. It is always better to close it the when you no longer need
it. DFOC (Dynamic File Open Close) can be done in program using the SET command
or linking to DFHEMTP.
DFOC BY LINK COMMAND
The application program links the master terminal program (DFHEMTP) and
passes CEMT parameter data required for the file open/close through COMMAREA.
Receiving the data, DFHEMTP program opens/closes the files specified, after which
the control is returned to the application program.
EXEC CICS LINK PROGRAM(DFHEMTP)
COMMAREA(WS-COMMAREA)
LENGTH(WS-LENGTH)
END-EXEC
The file open command is written into WS-COMMAREA and its length is
populated in WS-LENGTH.
File Open Command: SET DATASET(FCT-NAME) OPENED
DFOC BY SET COMMAND
EXEC CICS SET
DATASET(name) | FILE(name)
OPEN|CLOSED
END-EXEC
DFOC BY BATCH JOB
DFOC can be achieved from a batch job, which issues the CEMT as the data
for the JES Modify (F) command against the system console terminal.
// F cicsjob, CEMT SET DATASET(FCTNAME) OPENED
RANDOM ACCESS OF FILES
READ: EXEC CICS
READ FILE(FCT-NAME) [UPDATE]

INTO(WS-ITEM) | SET (ADDRESS OF LK-ITEM)
LENGTH(WS-LENGTH)
RIDFLD(WS-KEY)
KEYLENGTH(WS-KEY-LENGTH) GENERIC
SYSID(SYSTEM-NAME)
S1
S2
S3
S4
S5
S6
GTEQ | EQUAL
Page:190
S7
END-EXEC.
S1.
Reads the file and if update is intended then code UPDATE clause. This will get
exclusive access over the complete control interval where the specific record exists,
during the READ. Thus CICS ensures data integrity.
S2.
Read the file into working storage variable WS-ITEM. Alternatively, the
address of the record in the input-output area can be mapped to linkage section
variable by using SET command. Performance-wise SET is better than READ INTO.
VS COBOL2 supports ADDRESS of keyword and makes the code easier. In the prior
version COBOL, PARM list and SERVICE RELOAD should be used to achieve the same.
S3.
After the READ, the LENGTH of the record READ is updated into WS-LENGTH.
WS-LENGTH is the working storage half word binary item.
S4.
Key of the record to be read is moved into WS-KEY, a working storage item
and a part of record structure (WS-ITEM).
S5.
If partial key is used, then the length of the partial key should be moved to
WS-KEY-LENGTH and the keyword GENERIC should be coded. This is optional for fullkey read.
S6.
Remote system name where the request to be directed, is coded here. (1-4
character name)
S7.
EQUAL is default. GTEQ can be coded when you know the full key, but you are
not sure the record with that key exists in the file.
REWRITE
EXEC CICS REWRITE FILE(FCT-NAME)
FROM(WS-ITEM)
LENGTH(WS-LENGTH)
SYSID(SYSTEM-NAME)
END-EXEC.
To release the exclusive access acquired during the READ with UPDATE, the
record should be REWRITTEN using the above syntax. The parameters are selfexplanatory.
After you read record, if you feel that you dont want to update it, then inform
the same to CICS by issuing UNLOCK command so that CICS release the lock
acquired by you on the record.
EXEC CICS UNLOCK FILE(FCT-NAME) END-EXEC.
WRITE
The syntax of WRITE is same as REWRITE. There is a parameter RIDFLD with
Key-value is coded in the WRITE command. (Like S4 for READ command)
Page:191
When you want to add a set of records whose keys are in ascending order,
then qualify your first WRITE with another parameter MASSINSERT. This will get
exclusive control over the file and provide high performance during the mass insert.
If you use MASSINSERT, then you should issue UNLOCK to inform the
completion of your additions. CICS releases the file on UNLOCK.
DELETE
1.The record read using READ with UPDATE can be deleted using
EXEC CICS DELETE DATASET(FCT-NAME) END-EXEC.
2.The record in the file can directly be deleted by providing complete key.
EXEC CICS DELETE DATASET(FCT-NAME) RIDFLD(WS-KEY) END-EXEC.
3.Group of records can be deleted using partial key. After the deletion, the number of
records deleted is updated in the variable coded in NUMREC parameter. (WS-DELCOUNT)
EXEC CICS DELETE DATASET(FCT-NAME)
RIDFLD(WS-KEY)
KEYLEGNTH(WS-KEY-LENGTH) GENERIC
NUMREC(WS-DEL-COUNT)
END-EXEC
SEQUENTIAL READ:
Sequential access of VSAM file under CICS is called Browsing. It has FIVE
Commands associated with it.
STARTBR.
It establishes a position to start browsing.
EXEC CICS STARTBR FILE(FCT-NAME) RIDFLD(WS-KEY)
KEYLENGTH(WS-KEY-LENGTH) GENERIC
REQID(Integer-value)
SYSID(SYSTEM-NAME)
GTEQ | EQUAL
END-EXEC.
Meaning of the parameters is same as READ operation explanation. When you
want to do multiple browsing concurrently over the same file, then use REQID. The
first STARTBR can be coded with REQID(1) and the second STARTBR can be coded
with REQID(2).
One browse operation occupies one string of VSAM. If all VSAM strings are
exhausted for one VSAM file, then the other transactions will have to wait for one of
the strings to become free. So it is not recommended to use multiple browsing.
Instead, once the browsing is completed, using RESETBR set the position to another
place and start browsing. The syntax of RESETBR is same as STARTBR.
READNEXT and READPREV.
Page:192
EXEC CICS READNEXT| READPREV FILE(FCT-NAME) RIDFLD(WS-KEY)

INTO(WS-ITEM)
LENGTH(WS-LENGTH)
KEYLENGTH(WS-KEY-LENGTH)
REQID(INTEGER-VALUE)
SYSID(SYSTEM-NAME)
END-EXEC.
READNEXT is used to read the records in the forward direction from the
position established by STARTBR. READPREV is used to read the records in the
reverse direction. READPREV followed by READNEXT retrieves the same record once
again.
During the browse, if you want to skip set of records, then move the key of
the next record you intended to read to RIDFLD and then issue READNEXT. This is
called skip sequential read. Thus RIDFLD can be used as both input as well as output
field.
As VSAM files are variable length in most of the cases, WS-LENGTH should be
populated with maximum record length before issuing READ command.
KEYLENGTH(0) will position the cursor at the beginning of the file.
If you specify READPREV immediately after STARTBR, then you should code
key of a record that exists on the dataset. Otherwise NOTFND condition occurs on
READPREV.
ENDBR
It is used to terminate the current browse function.
EXEC CICS ENDBR FILE(FCT-NAME) REQID(INTEGER-VALUE) SYSID(SYSTEM-NAME)
END-EXEC.
UPDATE during BROWSE
Update and Browse are mutually exclusive functions. So if you want to update
a record during the browse operation, first issue ENDBR. Then give random read
using the key in hand with UPDATE option. REWRITE the record. Then issue STARTBR
with the key in hand and continue the browsing.
ESDS-BROWSING
Define a full-word binary item in working storage for the RBA of ESDS file.
Move LOW-VALUES or HIGH-VALUES to the field for the forward or reverse read
respectively. Issue STARTBR with RBA and EQUAL option. RIDFLD should point to
defined RBA field. Now issue READNEXT or READPREV for forward or REVERSE read
respectively.
GTEQ is invalid for ESDS.
MOVE LOW-VALUES TO ESDS-RBA
EXEC CICS STARTBR FILE(FCT-NAME) RIDFLD(ESDS-RBA) RBA EQUAL END-EXEC
MOVE 226 to WS-LENGTH.
EXEC CICS READNEXT FILE(FCT-NAME) INTO(WS-ITEM) RIDFLD(ESDS-RBA)
LENGTH(WS-LENGTH) RBA
END-EXEC.
Page:193
ESDS-WRITE
The syntax is same as KSDS WRITE command. Qualify the WRITE with RBA
option. The record will be appended to the file and the RBA of the record is placed
into RBA field mentioned in the WRITE command. (ESDS-RBA)
MOVE 225 TO WS-LENGTH
EXEC CICS WRITE FILE(FCT-NAME) RIDFLD(ESDS-RBA) LENGTH(WS-LENGTH) RBA
END-EXEC.
ESDS-RANDOM ACCESS
If you know the RBA value of the record you want to access, then you can
randomly access the ESDS file. The syntax is same as READ of KSDS file but qualify
it with RBA option. RIDFLD should point to full-word binary item pre-filled with RBA
of the record to be accessed.
RRDS ACCESS
RRDS file can be accessed using RRN in place of RBA and populating the
RIDFLD with RRN number. The field should be full word binary.
Alternate Index access
Register the path in FCT and use the path name as file name to read the file
randomly using alternate index. Please refer VSAM section of the book for
understanding what exactly PATH is.
As the alternate key need not be unique, DUPKEY condition is very common
during the alternate index usage and so it has to be properly handled.
Page:194
ASKTIME
EIBDATE and EIBTIME have the values at task initiation time. Upon the
completion of ASKTIME, these two fields are populated with current date and time.
EXEC CICS ASKTIME END-EXEC.
FORMATTIME
FORMATTIME is used to receive the information of date and time in various formats.
EXEC CICS FORMATTIME FORMAT-TYPE(data-area) END-EXEC
Format-type: YYDDD, YYMMDD, YYDDMM, MMDDYY, DDMMYY, DAYOFWEEK,
DAYOFMONTH, MONTHOFYEAR, YEAR, TIME, TIMESEP, and
DATESEP.
Example:
EXEC CICS
FORMATTIME MMDDYY(WS-DATE) DATESEP TIME(WS-TIME) TIMESEP
END-EXEC
WS-DATE contains the current date in MM/DD/YY (Default DATESEP is /)
WS-TIME contains the current time in HH:MM:SS (Default TIMESEP is :)
DELAY and SUSPEND
It is used to delay the processing of a task for a specified time interval or until
the specified time. If your task is doing some heavy CPU bounded work, it is a good
to place the DELAY command in order to allow the other tasks to proceed.
EXEC CICS DELAY INTERVAL(HHMMSS) | TIME(HHMMSS) END-EXEC.
SUSPEND is used to suspend a task. During the execution of the command,
the task will be suspended and the control will be given to other tasks with higher
priority. As soon as all higher priority tasks have been executed, control will be
returned to the suspended task.
EXEC CICS SUSPEND END-EXEC.
POST and WAIT EVENT
POST is used to request notification when the specific time has expired and
the WAIT EVENT command is used to wait for an event to occur. Usually these two
commands are used in pair, as alternative to the DELAY command.
EXEC CICS POST INTERVAL(HHMMSS) | TIME(HHMMSS)
SET (ADDRESS OF POST-ECB) END-EXEC.
END-EXEC
EXEC CICS WAIT EVENT ECAADDR(ADDRESS OF POST-ECB)
END-EXEC
CANCEL
It is used to cancel the interval control commands such as DELAY, POST and
START which have been issued. The interval commands to be cancelled are identified
using REQID.
EXEC CICS START REQID(STARTS) END-EXEC.
EXEC CICS CANCEL REQID(STARTS) END-EXEC.
Cancels all the starts issued.
Page:195
START and RETRIEVE

START command is used to start a transaction at the specified terminal and at
the specified time or interval. Optionally, data can be passed to the to-be-initiated
transaction.
EXEC CICS START
TRANSID(EMPC) TERMID(TST1)
S1
TIME(123000)|INTERVAL(003000)
S2
FROM(WS-ITEM) LENGTH(WS-LENGTH)
S3
RTRANSID(EMPD) RTERMID(TST0)
S4
QUEUE(EMPDET)
S5
END-EXEC.
S1.
EMPC is to be started at the terminal TST1
S2.
TIME and INTERVAL are mutually exclusive. If TIME(HHMMSS) is coded, then
the task start at the terminal at the specific time. In the above example it is 12:30.
If INTERVAL(HHMMSS) is coded then the task will start after the expiry of the
interval. In the example, the task starts after 30 minutes.
S3.
This program passes WS-ITEM of size WS-LENGTH to the task EMPC.
S4.
Transaction ID EMPD and Terminal Id TST0 are passed to EMPC. EMPC can
receive this information and may use for next triggering.
S5.
TSQ queue name can be passed from this transaction. EMPDET is passed in
our case.
The data passed by START command is received by the new transaction using
RETRIEVE command upon the expiry of START command.
EXEC CICS RETREIVE INTO(WS-ITME)| SET(ADDRESS OF LK-ITEM)
LENGTH(WS-LENGTH)
RTRANSID(WS-TRAN) RTERMID(WS-TERM)
QUEUE(WS-QUEUE)
END-EXEC.
ENQ and DEQ
ENQ gets the exclusive control over the resource and DEQ releases the
exclusive control acquired over the resource. As CICS takes care of exclusive access
over the resource whenever needed, it is advisable not to use it in application
programs.
But sometimes it may be needed in few cases like updating TSQ records or
getting access of the printer.
EXEC CICS ENQ/DEQ RESOURCE(QUEUE-NAME) END-EXEC.
SYNCPOINT and TWO-PHASE-COMMIT
The updates done by a task is automatically committed at task termination.
SYNCPOINT is used to commit portion of work completed without terminating the
task.
EXEC CICS SYNPOINT END-EXEC.
Page:196
During the program processing, if you want to roll back the changes made by
you, then you can code SYNPOINT with ROLLBACK. This will backed out all resource
modifications done since the last sync point.
EXEC CICS SYNCPOINT ROLLBACK END-EXEC.
SYNCPOINT in CICS-DB2 environment is called as two-phase commit.
The changes made in the resources that are directly under the control of CICS are
committed in the first phase and the changes made to DB2 environment are
committed in the second phase.
Queues
There are two types of queues available in CICS.
1.Temporary Storage Queue (TSQ)
2.Transient Data Queue (TDQ)
Their properties are listed below:
TSQ
No need to predefine anywhere. By default, TS queues are created in main

memory. So the content cannot be recovered after system crash. If you want
to recover, then you should define them in TST.
Records can be randomly accessed using ITEM option of READQ. READ is not
destructive.
Deletion of queue deletes all the records in it. Deletion of recoverable TSQ
should follow sync point before next WRITEQ.
Primarily used as scratch pad memory facility for any purpose.
Example: Page-up and Page-down logic, Passing huge data in between the
phases of transaction, Review and correction of multiple order entry screens.
TDQ
TDQ should be predefined in DCT.

READ is destructive and only sequential. Once the record is READ, it will be
logically deleted and cannot be read again.
Automatic Task Initiation: When number of records in the queue exceeds the
TRIGLEV defined in the DFHDCT entry of the queue, the transaction coded in
the TRANSID of DFHDCT is automatically triggered. This ATI is possible only
in TDQ.
There are two types of TDQ Intra and Extra partition. The DCT entry
identifies the type of queue from TYPE parameter.
Intra Partition Queues are used within a CICS region and DELETEQ deletes
all the records physically in the queue.
Extra partition Queues are used across the regions and systems. If you want
to pass some data from CICS to batch or receive data from Batch to CICS,
you should go for extra partition queue.
In the same program TDQ cannot be opened in both input and output mode.
Page:197
TSQ are preferred over TDQ for data passing. TDQ is used for batch interface
or on ATI requirement.
MRO and ISC

Within one processor, there could be more than one CICS region, each of
which runs independently under the same operating system. This is called Multi
region Operation. (MRO) .It is achieved by four inter communication facilities.
1.Function Request Shipping.
2.Asynchronous Processing.
3.Transaction routing.
4.Distributed Transaction Processing.
One CICS in one processor can communicate with other CICS in other
processor or other non-CICS system regardless of where the other processor
physically located. This is called Inter System Communication (ISC). It requires
sophisticated communication networks based on the System Network Architecture
(SNA).
MASTER TRANSACTIONS
CEMT:
This is Master Terminal transaction. It is menu driven and easy-to-use
transaction but due to its nature of manipulating the CICS environment, most of the
functions are restricted to application programmer or end user.
CEMT INQUIRE|SET|PERFORM
CEMT INQ TRAN|PRO|FILE Display information from PCT|PPT|FCT
CEMT INW TASK
Display the active running tasks in the region
CEMT SET command can be used to reset the values of PCT|PPT|FCT.
PERFORM has to be handled carefully. CEMT PERFORM SHUTDOWN will shut down
the entire CICS region.
Frequently used commands:
CEMT SET PR(PGM-NAME) NEW - To create a new copy of an application program
CEMT SET DA(file-name) CLO/OPEN
- To close/open a file from CICS.
CECI:
This is Command Interpreter transaction. CICS commands can be pre-tested
using this command before placing them into the program.
CEBR:
This is browse transaction. TSQ can be browsed using this transaction.
CEDF:
This is Diagnostic Facility transaction. If you want to debug your program step
by step, then before typing transaction ID type CEDF so that debug mode will be set.
It intercepts a transaction at initiation and termination of a program, before and after
execution of any EXEC CICS or EXEC SQL commands.
CMAC:
Page:198
This transaction is used to get ABEND codes and messages.

Different ways of initiating a transaction in CICS
1.By entering transaction identifier.
2.By START command.
3.By RETURN TRANSID
4.By registering the programs in PLT, they will be automatically initiated during CICS
startup.
5.Automatic Task Initiation of TDQ.
6.PF or PA keys could be defined in PCT to initiate a CICS transaction.
CICS Dump Analysis
Dump usually consists of Dump Header Area, TCA(User Area), TCA(System
Area), Trace Table, Transaction Storage and Program Storage.
At the top of the ABEND dump, there is a dump header area, which provides
ABEND code, TASK (the transaction ID), Program Status Word (PSW) and the
contents of registers at the time of ABEND.
PSW.
It is a double word field containing the status information at the time of
ABEND. 32nd bit says the addressing mode in use. If the addressing mode is 31 bits,
then the bit will be 1 and the bits 33 to 63 contains the next sequential instruction
(NSI), which would have been executed if the ABEND had not occurred.
First make sure that you have program compilation and link edited
SYSPRINTS of the program ABENDED and dump prints of the ABENDED transaction is
with you. Program should be compiled with LIST option.
Finding the statement caused the ABEND
1.Identify the address of next sequential instruction from the PSW.
2. From the program storage, identify the starting address of program storage.
3.Subtract the starting address of program storage from the NSI address identified in
the first step, you will get the displacement of the NSI from the beginning of program
storage.
4.Find the displacement of application program from the beginning of the load
module in the link edit list and subtract this value from the calculated NSI
displacement to get the real NSI displacement.
5.Find the statement in the compilation listing that corresponds to the OFFSET
calculated in the previous step. The statement one above is the statement that
caused the ABEND.
6.Analyse the statement for the cause of ABEND. This might have involved one or
more fields. Most of the times, you can easily come to conclusion the cause of the
Page:199
ABEND once you identified the statement. If you could not, then you may check the
value of the field at the time of ABEND as follows.
Finding the value of the field in error.
1.Find the Base Locator (BL), Displacement and length of the field from the MAP of
2. Find the register number corresponding to the base locator from the bottom of
3. Find the address of this register in the dump.
4. Add the displacement to the register of the address to get the address of the field.
5. Locate the address in the dump to get the value of the field.
CICS TABLE ENTRIES

FCT
All the VSAM files, PATH between base cluster and alternate index and any
BDAM datasets are to be registered in FCT to access them in the application
program.
DFHFCT
TYPE
=DATSET| FILE,
ACCMETH=BDAM | VSAM,
DATASET= name | FILE=name,
SERVREQ = (ADD, BROWSE, DELETE, READ, UPDATE),
[FILESTAT= (ENABLED|DISABLED, OPENED|CLOSED)],
[BUFND= n+1],
[BUFNI=n],
[STRNO=n],
[DSNAME=name],
[DISP=OLD | SHR],
[BASE=name],
ACCMETH defines data access method.

Name mentioned in the FILE parameter is the one with which we use in the
CICS file commands. The name should correspond to the DDNAME of the file
specified in the JCL of the CICS job itself. If DSNAME and DISP are coded, then JCL
entry is not needed.
SERVREQ lists the authorized input/output operations against the file. If other
than the specified services is requested in the file control command, then INVREQ
condition will occur.
FILESTAT indicates the initial file status when CICS initially starts. BUFND and
BUFNI indicate the number of data buffers and index buffers respectively. Default
value is one for BUFNI and two for BUFND. STRNO is the number of VSAM strings by
which concurrent access to the file is allowed.
If the file being registered is path, then BASE indicates the FCT name of the
base cluster.
Page:200
PPT
All the CICS application programs and BMS map sets must be registered in
PPT. If the program is not registered here, then the program is unrecognizable to
CICS.
DFHPPT
TYPE=ENTRY,
PROGRAM=name| MAPSET= name
[PGMLANG=(ASSEMBLER | COBOL | PLI)],
options..
ASSEMBLER is the default program language. All the map-sets are considered as
written in assembler.
PCT
The control information of all CICS transactions must be registered in
program control table (PCT).
DFHPCT
TYPE=ENTRY,
TRANSID=name,
TASKREQ=xxxx,
PROGRAM=name,
[DTIMOUT=mmss],
[RTIMOUT=mmss],
[RESTART= NO | YES],
[TRANSEC=1 | DECIMAL],
[DUMP=YES|NO],
Other options..
(PF1-PF24, PA1-PA3)
(1-64)
TRANSID
Transaction identifier name (1-4 characters)
PROGRAM
Name of the program associated with the transaction.
TASKREQ
- Key associated with the transaction initiation.
DTIMOUT
- Defines the timeout (waiting time) in case of deadlock.
RTIMOUT
- Defines the time limit within which the user has to respond. If
he didnt then the task will be cancelled.
RESTART
Automatic transaction is applied after the completion of the
transaction recovery from the abnormal termination of the transaction.
TRANSEC
Transaction security code.
DUMP
- Whether DUMP is to be taken in case of abnormal termination
of the transaction.
TCT
All the terminals, which are to be under CICS control, should be registered in
TCT. Terminal Control Program uses this table for identifying the terminals and
performs all input/output operations against these terminals.
DFHTCT
TYPE=ENTRY,
Page:201
ACCMETH=VTAM,
TRMIDNT=name,
TRMTYPE=type,
FEATURE=(UCTRAN,..)
Options..
TRMIDNT is (1-4 characters) terminal ID.
TRMTYPE is type of terminal. Ex. IBM3270.
FEATURE indicates the terminal services CICS offers. UCTRAN is used to
translate all
the lowercase characters to upper case characters.
DCT
TDQ control information is registered here. Destination Control Program uses

this table for identifying all TDQs and performs input/output operations against them.
Intra partition queue can be defined as follows.
DFHDCT
TYPE=INTRA,
DESTID=name,
[TRANSID=name],
[TRIGLEV=number],
[REUSE= YES|NO]
TYPE can be INTRA (Intra partition TDQ)
ATI:
TRANSID and TRIGLEV are used for automatic task initiation. If the number of
records in TDQ exceeded the number of records mentioned against TRIGLEV, then
transaction ID mentioned against TRANSID will be invoked automatically.
Once a record of intra partition TDQ is read by a transaction, the record is logically
removed from, the queue. If REUSE = YES is coded, then the space occupied by it
can be used by other records in future but the logically deleted records are not
recoverable.
Extra partitioned queue can be defined as follows.
DFHDCT
TYPE=EXTRA,
DESTID=name,
DSCNAME=name-2,
[OPEN=INITIAL|DEFERED]
DFHDCT
TYPE=SDSCI,
DSCNAME=name-2,
[TYPFILE=INPUT|OUTPUT|RDBACK]
Page:202
OPEN=INITIAL means the file will be open at the CICS start-up time and
DEFERED means the file will be closed until it is specifically opened by CEMT.
DSCNAME defines data control block and for one DSCNAME, a corresponding
DFHDCT entry must be made with TYPE=SDSCI and the same DSCNAME, which in
effect indicates DDNAME of extra partition dataset in JCL or CICS JOB itself.
TYPFILE indicates the file to be INPUT, OUTPUT or READ BACKWARD
(RDBACK)
How to submit JCL from a CICS program?
JCL can be submitted using the SPOOLOPEN, SPOOLWRITE and SPOOLCLOSE
commands.
First you must open a spool for output using SPOOLOPEN

EXEC CICS
SPOOLOPEN OUTPUT NODE('*') USERID(INTRDR)
TOKEN(report_token) RESP(response field)
END-EXEC.
NODE('*') specifies a destination node to send the JCL to. USERID is set to
the name of the internal reader INTRDR. A unique token will be allocated by CICS
when the SPOOLOPEN is executed and placed in the field you specify using
TOKEN(report_token). The token will be used as a sending field on subsequent
commands. The token will be 8 bytes long.
To write each line of the job to the spool use the SPOOLWRITE command.
EXEC CICS
SPOOLWRITE FROM(io_area) TOKEN(report_token)
RESP(resonse_field)
END-EXEC.
The "io_area" field should be the name of a data item containing the line of
JCL. The "report_token" field should be the same as the 8 byte token returned from
SPOOLOPEN. An end of job statement ('//' or '/*EOF') should be written as the last
line.
Finally you must close the spool using SPOOLCLOSE. (Note if you do not
explicitly close the spool it will be closed when the transaction terminates,
However it is good practice to close the spool explicitly)
EXEC CICS
SPOOLCLOSE TOKEN(report_token) RESP(response_field)
END-EXEC.
Again the "report_token" field must be the same as the one allocated at
SPOOLOPEN.
Notes:
The RESP option should be coded on all the Commands and it is
recommended that you also code RESP2, because additional information is
returned for some exception conditions in this field.
DCT entries and JCL DD statements are not needed when using this method.
Page:203
Note that in order to use this method DFHSIT SPOOL=YES must be coded in
the CICS System Initialization Table. Check with your friendly local SysProg if
you are unsure.
Under OS/VS COBOL the SPOOLWRITE command had to have FLENGTH
specified. FLENGTH specifies the length of the data being written in a fullword
binary field. This field is optional in newer versions, if it is omitted then the
length is assumed to be the length of the data item (io-area) specified in
FROM.
Be aware of the performance considerations for writing to the spool, IBM says
"Transactions that process SYSOUT data sets larger than 1000 records, either for
INPUT or for OUTPUT, are likely to have a performance impact on the rest of CICS".
(Source: CICS/ESA V4R1 Application Programming Reference SC33-1170-00)
Page:204

Complete Mainframe

Uploaded by

Copyright:

Available Formats

Complete Mainframe

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Complete Mainframe

Uploaded by

Copyright:

Available Formats

What are the two types of commands that can be entered in the edit/view panel of ISPF editor?

What are the two types of commands that can be entered in the edit/view panel of ISPF editor?

What does the RECOVERY option do in ISPF?

What does the RECOVERY option do in ISPF?

Mainframe Refresher Part-1

Mainframe Refresher Part-1

Mainframe Refresher Part-1

unavailable warning message.

Locates the First/Last Musa in the dataset/member

Locates the Next/Previous Musa with respect to the

Note: ALL/FIRST/NEXT/PREV/LAST are mutually exclusive

Look for Musa . Word is a string followed and prefixed by

Look for Musa as a prefix or suffix with some other

DEL ALL X/NX

Note: Change command is nothing but find and replace. So

Mainframe Refresher Part-1

Displays any numeric

Displays any upper-case

Displays any non-blank

C P 73 80 ALL changes all non-blank characters in 73-80 to spaces.

Mainframe Refresher Part-1

Move the block of lines

Repeats the block of lines. RR 10

Exclude this line. X10 excludes

Shift the line content by n

Copy the block of lines

Insert 10 lines. Once enter is

Mainframe Refresher Part-1

the terminal. (Type data without

Overlay command indicates the

O10 overlaid the first 10 lines.

Mainframe Refresher Part-1

datasets start with BPMAIN.TEST to BPMAIN.UNIT.

RENAME SMSXL86.TEST(MEM1) MEM2 renames mem1 as

Mainframe Refresher Part-1

Mainframe Refresher Part-1

Job Control Language

Mainframe Refresher Part-1

Mainframe Refresher Part-1

Job Entry Subsystem (JES)

Once the work to be done is defined in JCL, it can be submitted to the

Mainframe Refresher Part-1

Programmer name or program functionality or group can be mentioned. It is

1. CLASS is coded with single alphanumeric character (A-Z, 0-9).

Syntax: PRTY=N (N can be 0 15).

Syntax: MSGLEVEL=(X,Y) (X can be 0-2 & Y can be 0-1)

Mainframe Refresher Part-1

MSGCLASS (Installation Defined)

Syntax: MSGCLASS=X (X can be A-Z, 0-9).

Syntax: REGION={xK | yM} (x can be 1-2096128 & y can be 1-2047).

It is used to request special job processing.

Mainframe Refresher Part-1

It defines the Step and Step level information to the system.

Mainframe Refresher Part-1

Code the Program name or PROC name to be executed. (1-8 characters)

1. It is used to pass variable information to the processing program, executed by this

It is used for conditionally executing one or more steps. Nesting is possible up

Mainframe Refresher Part-1

the name SMSXL86.EMPLOYEE.DETAILS, one of the programs (EMPPGM) in the

Dataset does not exist. It will be created in this step