DFSORT/ICETOOL question

John W Perry · New User Joined: 29 Jan 2014 Posts: 8 Location: USA

Hi,

I am attempting to:

1. Sort and split 'IN' file into the two output files as shown.
2. Write the existing header to each of the two output files.
3. Plug a fixed value of 'ABC' into the header of OT1.
4. Plug a fixed value of 'XYZ' into the header of OT2.

IN:
HEADER 000
1111111 B06
2222222 B05 006
3333333 B04 006
4444444 B03
5555555 B02 006
6666666 B01

OT1:
HEADER ABC
6666666 B01
4444444 B03
1111111 B06

OT2:
HEADER XYZ
5555555 B02 006
3333333 B04 006
2222222 B05 006

The JCL for my initial attempt is below, but discovered duplicate OUTREC statements are not allowed.

I was able to accomplish my goal - but with two separate steps. Just wondering if there was a way to get it done in one step.

//SPLIT EXEC PGM=ICETOOL
//IN DD DSN=&HLQ..IN,DISP=SHR
//OT1 DD DSN=&HLQ..OT1,
// DISP=(NEW,CATLG),
// SPACE=(TRK,(1,1),RLSE)
//OT2 DD DSN=&HLQ..OT2,
// DISP=(NEW,CATLG),
// SPACE=(TRK,(1,1),RLSE)
//TOOLMSG DD SYSOUT=*
//DFSMSG DD SYSOUT=*
//TOOLIN DD *
SORT FROM(IN) USING(OUTA)
/*
//OUTACNTL DD *
SORT FIELDS=(9,3,CH,A,1,7,CH,A)
OUTFILE FNAMES=(OT1),
INCLUDE=(1,6,CH,EQ,C'HEADER',OR,13,3,CH,EQ,C' ')
OUTREC IFTHEN=(WHEN=(1,6,CH,EQ,C'HEADER'),
OVERLAY=(17:C'ABC'))
OUTFILE FNAMES=(OT2),
INCLUDE=(1,6,CH,EQ,C'HEADER',OR,13,3,CH,EQ,C'006')
OUTREC IFTHEN=(WHEN=(1,6,CH,EQ,C'HEADER'),
OVERLAY=(17:C'XYZ'))
/*

Bill Woodger · Posted: Sat May 31, 2014 3:27 am

You have fallen into the trap of becoming confused by OUTREC and OUTREC on OUTFIL. Two different things.

On OUTFIL, use BUILD instead of OUTREC, and you need a continuation comma as the last of each line before each BUILD.

John W Perry · New User Joined: 29 Jan 2014 Posts: 8 Location: USA

Thanks for your help Bill,

The use of BUILD worked fine.

I've also added a COPY step to produce two more output files in the original sort order. The process works OK but I was wondering if it could be accomplished with just one pass on the input file.

Bill Woodger · Posted: Thu Jun 05, 2014 4:32 am

Sorry John, just found this in the HomeWork part of the forum, which isn't somewhere questions appear from (I thought it had been deleted after my answer).

If you want the output sorted and in the original order, you will have to do it in separate operations, as you have done already. OUTFIL (or after OUTREC, if no OUTFIL present) is where data is written, and the input is long gone by that point.

The one thing you could try is to use an E15 exit, which can be written in COBOL. You will be presented with each record from the input, which you can then format and write to new files, from the COBOL program. Whether you data and the number of files you require means that this will use fewer resources is going to be down to experimentation.