OBJECTIVE:
1) To through some light on Sort card.
2) To extract particular number of records from a sequential file, which involves condition like select some records, then skip few records, then again select records and so on.
SORT UTILITY: SYNCSORT, DFSORT
We will discuss above thing by taking an example,
REQUIREMENT:
From a sequential file suppose we need to extract say first 5 records, skip next 4 records,
and again select next 5 records.
(Think this condition from the following point of view says if we have 100 conditions then what?)
For simplicity only have taken only two conditions. Assume that Input file has 20 records.
SOLUTIONS:
1.Using SKIPREC, STOPAFT parameters in SORT card.
2.Using OUTFIL in sort card.
3.Using OUTFIL with SAVE.
FIRST SOLUTIONS,
JCL would go like this,
FIRST STEP IN JCL would contain following sort card
//SYSIN DD *
SORT FIELDS=COPY,
STOPAFT=5
END
/*
Above sort card will select first five records from input file.
SECOND STEP IN JCL will contain following sort card,
a) To skip next 5 +4 =9 records,
b) To copy next five records starting from 10 Th record,
//SYSIN DD *
SORT FIELDS=COPY,
SKIPREC=9,
STOPAFT=5
END
THIRD STEP:
This step is there to finally merge the files, which are created in previous two steps. Output of this merge step will be our required file.
DRAWBACK:
Here the number of condition is 2, so we needed only two steps.
Think about! What will you do if we have 100 conditions? Do we write?
100 STEPS? Surely not, it would be cumbersome.
To avoid this we have second approach,
SECOND SOLUTION:
Using OUTFIL in sort card.
Objective:
1) To solve above problem with minimum number of steps.
2) To have only one sort card for all condition.
3) And one final merge step to merge intermediate files.
Sort card that can be used would be like this,
FIRST STEP IN JCL will contain following sort card
//SYSIN DD *
SORT FIELDS=COPY
OUTFIL FILES=1,STARTREC=1,ENDREC=5
OUTFIL FILES=2,STARTREC=10,ENDREC=14
END
/*
Above sort card will do following things,
1. 1. One file will create which will have first five records.
2. 2. Second file will created which will have next five records starting from 10 Th record
Make sure that you give following logical name of those intermediate files, which are created as SORTOF1,SORTOF2, SORTOF3 and so on.
SECOND STEP IN JCL,
Its a merge step, it will merge all files which are created above.
CONCLUSION:
Even 100 conditions are there, we would need only 2 steps to extract desired records from file.
DRAWBACK:
To avoid this we have third solution,
THIRD SOLUTION:
It’s a general solution independent of number of conditions. Also it can do in SINGLE STEP.
JCL would be like this,
//N010053A JOB (1L17L120),XXXX.III,
...
...
...
...
//STEP11 EXEC PGM=IERRCO00,REGION=650K,RD=R
//* PARM='VSCORE=5500K,MSG=AP,BMSG'
//SORTLIB DD DSN=SYS1.SORTLIB,DISP=SHR
//MODLIB DD DSN=SYS1.LOADLIB,DISP=SHR
//SYSOUT DD SYSOUT=(,),OUTPUT=(*.TOPAPER,*.TOHOLDQ)
//SYSUDUMP DD SYSOUT=(,),OUTPUT=(*.TOPAPER,*.TOHOLDQ)
//SORTIN DD DSN=T145.INPUT.FILE,
// DISP=SHR
//SORTOF1 DD DSN=&&TEMP1
//SORTOF2 DD DSN=&&TEMP2
//SORTOF3 DD DSN=T145.OUTPUT.FILE,
// DISP=(NEW,CATLG,DELETE),UNIT=SYSALLDA,
// SPACE=(8048,(545,2178),RLSE),
// DCB=(BLKSIZE=0,LRECL=53,RECFM=FB)
//SORTWK01 DD SPACE=(CYL,(502,101),RLSE),UNIT=DASD
//SORTWK02 DD SPACE=(CYL,(502,101),RLSE),UNIT=DASD
//SORTWK03 DD SPACE=(CYL,(502,101),RLSE),UNIT=DASD
//SORTWK04 DD SPACE=(CYL,(502,101),RLSE),UNIT=DASD
//DUMPDD DD SYSOUT=*
//DSPLD DD SYSOUT=*
//SYSIN DD *
SORT FIELDS=COPY
OUTFIL FILES=1,STARTREC=6,ENDREC=9
OUTFIL FILES=2,STARTREC=15,ENDREC=20
OUTFIL FILES=3,SAVE
END
/*
//*
Few points regarding this JCL,
1. 1. Use of SAVE in sort card.
2. 2. First temporary file will have four records after skipping first five records.
3. 3. Second temporary file will have records starting from 15th record till the last record.
4. 4. Here the catch is file 1 and file 2 will have all those records, which we don’t want.
5. 5. So file 3 ( T145.OUTPUT.FILE ) will have all first five records ,skipping next four records, again extracting next five records.
6. 6. All intermediate files will be deleted at end of JOB.
CONCLUSION:
Thus, only one step is required to extract records from file. Also no merge step is required now. One more thing, in all of the above sortcard we have mentioned so far, we can cover conditions like INCLUDE, OMIT etc also.
