Split a Dataset into multiple Members

yianis · New User Joined: 14 Aug 2006 Posts: 45

I have a job (DFSORT) for now and I read in a Dataset File and want to split it up into multiple output Members.

Here is a sample of my job:

Douglas Wilder · Posted: Fri Nov 16, 2007 9:03 pm

superk · Posted: Fri Nov 16, 2007 9:10 pm

The job seems fine. The B14 makes me suspect a problem with the PDS (not enough space, not enough directory blocks, etc.).

yianis · New User Joined: 14 Aug 2006 Posts: 45

Actually, i increased the space etc.. and the error is IEC143I 213-30
It's because I am trying to update/create multiple members in the dataset and is causing a problem

It won't allow me to have multiple output members within a single step.

I don't want to split this job into n-multiple steps with each step having a single input and single output...I'd prefer to do it in 1 step but it seems you cannot do this...

superk · Posted: Fri Nov 16, 2007 9:21 pm

I forgot, you're right, you can't do that. Maybe you'll find your answer using FileAid, unless a combination of SORT and IEBUPDTE would work for you?

I'll move this post over to the FileAid forum.

yianis · New User Joined: 14 Aug 2006 Posts: 45

Yeah I know.. that's why I had it originally posted in the JCL forum...

Craq Giegerich · Posted: Fri Nov 16, 2007 10:22 pm

yianis · New User Joined: 14 Aug 2006 Posts: 45

I could do the same thing with the DFSORT (multiple steps), but I am trying to avoid the creation of multiple steps and would like to do it in a single step. The actual jobb will probably have 100 files to create and I want to avoid 100 steps...Thx for the suggestion

Douglas Wilder · Posted: Fri Nov 16, 2007 11:56 pm

You could use separate PS datasets (with the last node different).
Or use a PDSE instead of a PDS.

lcmontanez · New User Joined: 19 Jun 2007 Posts: 50 Location: Chicago

you can also create a 1 step proc that is run 100 times.....

gcicchet · Posted: Tue Nov 20, 2007 7:25 am

Hi,
you can get away with it in 2 steps.

See below

Gerry [/code]
//STEP0001 EXEC PGM=ICEMAN
//SYSOUT DD SYSOUT=*
//SORTIN DD *
003
004
005
//*
//OUT001 DD DSN=&&O001,DISP=(,PASS,DELETE),UNIT=SYSDA,SPACE=(TRK,(1))
//OUT002 DD DSN=&&O002,DISP=(,PASS,DELETE),UNIT=SYSDA,SPACE=(TRK,(1))
//OUT003 DD DSN=&&O003,DISP=(,PASS,DELETE),UNIT=SYSDA,SPACE=(TRK,(1))
//SYSIN DD *
OPTION COPY
OUTFIL INCLUDE=(01,03,CH,EQ,C'003'),FNAMES=OUT001
OUTFIL INCLUDE=(01,03,CH,EQ,C'004'),FNAMES=OUT002
OUTFIL INCLUDE=(01,03,CH,EQ,C'005'),FNAMES=OUT003
//*
//**********************************************************************
//*
//REPRO EXEC PGM=IDCAMS
//SYSPRINT DD SYSOUT=*
//FILEIN1 DD DSN=&&O001,DISP=(OLD,DELETE)
//FILEIN2 DD DSN=&&O002,DISP=(OLD,DELETE)
//FILEIN3 DD DSN=&&O003,DISP=(OLD,DELETE)
/*
//FILEOUT1 DD DSN=PDS(A),DISP=SHR
//FILEOUT2 DD DSN=PDS(B),DISP=SHR
//FILEOUT3 DD DSN=PDS(C),DISP=SHR
//SYSIN DD *
REPRO INFILE(FILEIN1) OUTFILE(FILEOUT1)
REPRO INFILE(FILEIN2) OUTFILE(FILEOUT2)
REPRO INFILE(FILEIN3) OUTFILE(FILEOUT3)

yianis · New User Joined: 14 Aug 2006 Posts: 45

Thanks... I actually created a dataset with the type as LIBRARY (i.e. PDSE) and it worked...I will also keep your suggestion in mind too.