IBM Mainframe Forum

by **Laksh** » Tue Jan 08, 2008 12:41 pm

My input file has the first seven bytes as the sort tag. I need a sort that would give me two output files - One with unique records with the first seven bytes and the count of such records. The second should be the same as my input file with the first seven bytes removed. Is it possible to write such a Sort?

My input file:
ABCDEFG jkhfjdkhmhgjldhglkdgh
ACDFGRE ljglkrgjrklgjrklgjrklgjkl
BFGRSFF dgdgdkglrgmlgmlglgklkl
ABCDEFG jfkfjkjlmhnfdjknfhssjs
ASDFREG jklejf agfwdhdsbndmfj
ACDFGRE hwitbbcvmznhyurhwin

My output files should be:
Output 1:
ABCDEFG 2
ACDFGRE 2
ASDFREG 1
BFGRSFF 1

Output 2: This can be sorted or may be written in the same order as in the input file.
jkhfjdkhmhgjldhglkdgh
ljglkrgjrklgjrklgjrklgjkl
dgdgdkglrgmlgmlglgklkl
jfkfjkjlmhnfdjknfhssjs
jklejf agfwdhdsbndmfj
hwitbbcvmznhyurhwin

Thanks in advance,
Laksh

by **Alissa Margulies** » Fri Feb 29, 2008 11:26 pm

This can be accomplished in a single pass of the data using PIPESORT, which is an add-on product to SyncSort for z/OS. Assuming that you do NOT have access to PIPESORT, here is a SyncSort application that should provide the desired results:

Select all

//STEP1 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTOUT DD SYSOUT=*
//SORTIN DD *
ABCDEFG JKHFJDKHMHGJLDHGLKDGH
ACDFGRE LJGLKRGJRKLGJRKLGJRKLGJKL
BFGRSFF DGDGDKGLRGMLGMLGLGKLKL
ABCDEFG JFKFJKJLMHNFDJKNFHSSJS
ASDFREG JKLEJF AGFWDHDSBNDMFJ
ACDFGRE HWITBBCVMZNHYURHWIN
//SYSIN DD *
INREC FIELDS=(1,16)
SORT FIELDS=(1,7,ZD,A)
OUTFIL REMOVECC,NODETAIL,
SECTIONS=(1,7,TRAILER3=(1:1,7,9:COUNT))
/*
//STEP2 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTOUT DD SYSOUT=*
//SORTIN DD *
ABCDEFG JKHFJDKHMHGJLDHGLKDGH
ACDFGRE LJGLKRGJRKLGJRKLGJRKLGJKL
BFGRSFF DGDGDKGLRGMLGMLGLGKLKL
ABCDEFG JFKFJKJLMHNFDJKNFHSSJS
ASDFREG JKLEJF AGFWDHDSBNDMFJ
ACDFGRE HWITBBCVMZNHYURHWIN
//SYSIN DD *
SORT FIELDS=COPY
OUTREC FIELDS=(9,25)
/*

by **Alissa Margulies** » Sat Mar 01, 2008 1:08 am

For the first step, you can also specify:

Select all

//SYSIN DD *
SORT FIELDS=(1,7,CH,A)
INREC OVERLAY=(1,7,9:C'1')
SUM FIELDS=(9,1,ZD)
OUTREC FIELDS=(1,9)

by **arunprasad.k** » Sat Mar 01, 2008 1:18 pm

Laksh, How about this JCL in single pass?

Select all

//S01 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//OUT1 DD SYSOUT=*
//OUT2 DD SYSOUT=*
//SORTIN DD *
ABCDEFG JKHFJDKHMHGJLDHGLKDGH
ACDFGRE LJGLKRGJRKLGJRKLGJRKLGJKL
BFGRSFF DGDGDKGLRGMLGMLGLGKLKL
ABCDEFG JFKFJKJLMHNFDJKNFHSSJS
ASDFREG JKLEJF AGFWDHDSBNDMFJ
ACDFGRE HWITBBCVMZNHYURHWIN
//SYSIN DD *
SORT FIELDS=(1,7,ZD,A)
OUTFIL FNAMES=OUT1,REMOVECC,NODETAIL,
SECTIONS=(1,7,TRAILER3=(1:1,7,9:COUNT))
OUTFIL FNAMES=OUT2,OUTREC=(1:9,25)
/*

Output 2: This can be sorted or may be written in the same order as in the input file.

My JCL will give the O/P in sorted order.

Post if you have any question.

Arun.

by **Alissa Margulies** » Mon Mar 03, 2008 11:12 pm

arunprasad.k wrote:My JCL will give the O/P in sorted order.

Sorry, I misread the initial post and thought the second output had to be the same order of the original input file.

IBM Mainframe Forum

Sort reqd to sum the records & also to reformat the records.

Sort reqd to sum the records & also to reformat the records.

Re: Sort reqd to sum the records & also to reformat the records.

Re: Sort reqd to sum the records & also to reformat the records.

Re: Sort reqd to sum the records & also to reformat the records.

Re: Sort reqd to sum the records & also to reformat the records.