IBM Mainframe Forum

by **dja2** » Tue Oct 02, 2012 8:45 pm

As you can see from the attached code, I am creating two datasets, with data from different parts of the input record.

Is it possible to de-duplicate the records in the output dataset in the same SORT step? (Code follows).
SORT FIELDS=COPY
OUTFIL FNAMES=CONNID,
BUILD=(1,6,C' ',264,4),
OUTFIL FNAMES=INTCSTID,
BUILD=(1,6,C' ',278,4)

I can, (and have), de-duplicated the datasets in a separate, following, SORT step. I just wondered if it was possible to do it all in one step.

by **BillyBoyo** » Tue Oct 02, 2012 8:58 pm

If you show all the code for both the steps then we might have suggestions.

by **dick scherrer** » Tue Oct 02, 2012 9:02 pm

Hello,

Suggest you post a bit of sample input data and the output you want when this input is processed. The records do Not need to be full-width, just enough to show what you have and want to do with it.

If i understand what you are asking, yes, this can be done in a single step.

by **skolusu** » Tue Oct 02, 2012 10:36 pm

dja2,

If your intention is write out 2 records for every record in the input file with contents from pos 264 and 278, you can use the following DFSORT JCL. You need to use the operator "/" to split the records into 2.

Select all

//STEP0100 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTIN DD DSN=Your input file,DISP=SHR
//SORTOUT DD SYSOUT=*
//SYSIN DD *
SORT FIELDS=COPY
OUTFIL BUILD=(1,6,X,264,4,/,
1,6,X,278,4)
//*

by **dja2** » Wed Oct 03, 2012 1:30 pm

Apologies for not giving enough information.

This is a sample of the input data

Select all

=COLS> ----+----1----+----2----+----3----+----4----+----5--
000008 WAVE 3 15202510252825 15202510252825 KEY BB RD
000009 WAVE 3 15202510283461 15202510283461 KEY BB RD
000010 WAVE 3 15202510283461 15202510283461 KEY BB RD
000011 WAVE 3 15202510283488 15202510283488 KEY BB RD
000012 WAVE 3 15202510283488 15202510283488 KEY BB RD

Using this sample, from this input, I create three output files, every file contains columns 1 to 6 ("WAVE 3").
Output dataset 1 also contains columns 8 to 27.
Output dataset 2 also contains columns 28 to 41.
Output dataset 3 also contains columns 43 to 56.

So, each output dataset will contain 5 records.

Output datasets 1 and 2 will contain 3 distinct records.
Output dataset 3 will contain 1 distinct record.

To concentrate on output dataset 1, this will contain;

Select all

=COLS> ----+----1----+----2----+--
000001 WAVE 3 15202510252825
000002 WAVE 3 15202510283461
000003 WAVE 3 15202510283461
000004 WAVE 3 15202510283488
000005 WAVE 3 15202510283488

My question is, how do I remove the duplicate records from the output datasets, to give the following ?

Select all

=COLS> ----+----1----+----2----+--
000001 WAVE 3 15202510252825
000002 WAVE 3 15202510283461
000003 WAVE 3 15202510283488

A portion of my JCL follows

Select all

//STEP100 EXEC PGM=SORT,COND=(0,NE)
//SORTIN DD DSN=XXXX.SEQ.WAVE4,DISP=OLD
//CONNID DD DSN=XXXX.SEQ.CONNID.UNSRTD,
// DISP=(NEW,CATLG,DELETE),
// DCB=(LRECL=27,RECFM=FB),
// DSORG=PS,
// SPACE=(CYL,(1,20),RLSE)
//INTCSTID DD DSN=XXXX.SEQ.INTCSTID.UNSRTD,
// DISP=(NEW,CATLG,DELETE),
// DCB=(LRECL=21,RECFM=FB),
// DSORG=PS,
// SPACE=(CYL,(1,20),RLSE)
//CIN DD DSN=XXXX.SEQ.CIN.UNSRTD,
// DISP=(NEW,CATLG,DELETE),
// DCB=(LRECL=21,RECFM=FB),
// DSORG=PS,
// SPACE=(CYL,(1,20),RLSE)
//SORTMSG DD SYSOUT=*
//SYSOUT DD SYSOUT=*
//SYSPRINT DD SYSOUT=*
//SYSUDUMP DD SYSOUT=*
//SYSIN DD *
SORT FIELDS=COPY
OUTFIL FNAMES=CONNID,
BUILD=(1,27),
OUTFIL FNAMES=INTCSTID,
BUILD=(1,7,28,14)
OUTFIL FNAMES=CIN,
BUILD=(1,7,43,14)
//*

Above corrected to Output dataset 3 also contains columns 43 to 56.

by **NicC** » Wed Oct 03, 2012 1:37 pm

Do you mean:
Output dataset 3 also contains columns 43 to 56

by **dja2** » Wed Oct 03, 2012 2:01 pm

NicC

Apologies - typo - I should have written Output dataset 3.

by **BillyBoyo** » Wed Oct 03, 2012 2:14 pm

If the duplicates can only be contiguous, then have a look at the "reporting functions" on OUTFIL. SECTIONS=(start,lengh with a TRAILERn will do a "consolidation". With REMOVECC (don't include the printer control-code) and NODETAIL (exclude detail records) you could get what you want.

If the "key fields" are not contiguous, you are going to have to SORT in a separate step.

by **dja2** » Wed Oct 03, 2012 2:25 pm

BillyBoyo

Thanks very much. The "Key Fields" will not necessarily be contiguous, so it looks as though another SORT will have to be done.

by **skolusu** » Wed Oct 03, 2012 9:39 pm

dja2,

If your intention is to get the unique records in columns 8 to 27 and columns 43 to 56, then the following DFSORT JCL will give you the desired results. The trick here is to use the same file twice and match on unique which we create using JNF1/JNF2

Select all

//STEP0100 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//INA DD *
WAVE 3 15202510252825 15202510252825 KEY BB RD
WAVE 3 15202510283461 15202510283461 KEY BB RD
WAVE 3 15202510283461 15202510283461 KEY BB RD
WAVE 3 15202510283488 15202510283488 KEY BB RD
WAVE 3 15202510283488 15202510283488 KEY BB RD
//INB DD *
WAVE 3 15202510252825 15202510252825 KEY BB RD
WAVE 3 15202510283461 15202510283461 KEY BB RD
WAVE 3 15202510283461 15202510283461 KEY BB RD
WAVE 3 15202510283488 15202510283488 KEY BB RD
WAVE 3 15202510283488 15202510283488 KEY BB RD
//CONNID DD SYSOUT=*
//CIN DD SYSOUT=*
//SYSIN DD *
JOINKEYS F1=INA,FIELDS=(29,8,A),SORTED,NOSEQCK
JOINKEYS F2=INB,FIELDS=(29,8,A),SORTED,NOSEQCK
JOIN UNPAIRED
REFORMAT FIELDS=(F1:1,36,F2:1,36,?)
INREC IFOUTLEN=28,IFTHEN=(WHEN=(73,1,CH,EQ,C'2'),BUILD=(37,28))
SORT FIELDS=(8,21,CH,A)
SUM FIELDS=NONE
OUTFIL FNAMES=CONNID,INCLUDE=(28,1,CH,EQ,C'A'),BUILD=(1,27)
OUTFIL FNAMES=CIN,INCLUDE=(28,1,CH,EQ,C'B'),BUILD=(1,21)
//*
//JNF1CNTL DD *
INREC BUILD=(1,27,C'A',SEQNUM,8,ZD,START=1,INCR=2)
//*
//JNF2CNTL DD *
INREC BUILD=(1,7,43,14,28:C'B',SEQNUM,8,ZD,START=2,INCR=2)
//*

IBM Mainframe Forum

De-duplicating records in the outfil

De-duplicating records in the outfil

Re: De-duplicating records in the outfil

Re: De-duplicating records in the outfil

Re: De-duplicating records in the outfil

Re: De-duplicating records in the outfil

Re: De-duplicating records in the outfil

Re: De-duplicating records in the outfil

Re: De-duplicating records in the outfil

Re: De-duplicating records in the outfil

Re: De-duplicating records in the outfil