IBM Mainframe Forum

by **akanchu** » Sat Aug 08, 2009 7:26 am

Hi,
Checking to see if this is possible with SORT/ICETOOL.
I have a file that has records in the following format.

HEAD MCNTRXX 0807200916301242 0000000001
HEAD CNTARXX 0807200916301242 0000000002010766100
10000085000A010766100
30 0003500001
70 14 GRABLES CIR
HEAD CNTBRXX 0807200916301242 000000000351800760873
10000085000A51800760873
30 0003500001
70 3 FENDER CT
HEAD CNTARXX 0807200916301242 0000000010010768900
10000085000A010768900
23 0003500001
73 56 JILLCREST RD

.
.
.

HEAD MCNTRXX 0807200916301242 0000000013

The first record is the header of the file that needs to be skipped.
The second record is the header for the group.
So the records :
HEAD CNTARXX 0807200916301242 0000000002010766100
10000085000A010766100
30 0003500001
70 14 GRABLES CIR
are 1 set.
simiarly others.

The last record is the trailer of the file, with a count in it, that can be skipped.

The task here is to split the file into 2 files so that records with its header and set of records should be written to 1 file.. and other group to 2nd file.
Last record should be the trailer with the count of that file.

The first 12 bytes is the key - in this case "HEAD CNTARXX" and "HEAD CNTBRXX"

So the result should be :
File 1 :

Select all

HEAD CNTARXX 0807200916301242 0000000006010766100
10000085000A010766100
30 0003500001
70 14 GRABLES CIR
HEAD CNTARXX 0807200916301242 0000000010010768900
10000085000A010768900
23 0003500001
73 56 JILLCREST RD
HEAD CNTARXX 0807200916301242 0000000002

File 2 :

Select all

HEAD CNTBRXX 0807200916301242 000000000351800760873
10000085000A51800760873
30 0003500001
70 3 FENDER CT

HEAD CNTBRXX 0807200916301242 0000000001

Hope I have drafted my question properly.
Let me know if anyone needs more clarification.

Thanks
Aka

by **Frank Yaeger** » Mon Aug 10, 2009 9:32 pm

Let me know if anyone needs more clarification.

Yes:

What is the RECFM and LRECL of the input file?

What is the starting position, length and format of each relevant field?

The first record can just be ignored - right?

Where does the following field in the trailer record come from

0807200916301242

Is it just a hardcoded string or is it copied from one of the input records (which one)?

It appears the count in the trailer record is the count of the number of groups (e.g. the number of CNTARXX groups for output file1) - is that correct?

by **akanchu** » Tue Aug 11, 2009 12:25 am

Here are the answers :

What is the RECFM and LRECL of the input file?
RECFM is FB and LRECL is 3000 bytes
What is the starting position, length and format of each relevant field?
the key starts from 1 to 13 bytes.
The first record can just be ignored - right?
Yes it can be ignored
Where does the following field in the trailer record come from

0807200916301242

Is it just a hardcoded string or is it copied from one of the input records (which one)?
it is hardcoded same pos as the other records in the key record

It appears the count in the trailer record is the count of the number of groups (e.g. the number of CNTARXX groups for output file1) - is that correct?
Yes your understanding is correct, it is the count of number of groups

thanks.

by **Frank Yaeger** » Tue Aug 11, 2009 1:33 am

Here's a DFSORT job that will do what you asked for:

Select all

//S1 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTDIAG DD DUMMY
//SORTIN DD DSN=... input file (FB/3000)
//OUT1 DD DSN=... output file1 (FB/3000)
//OUT2 DD DSN=... output file2 (FB/3000)
//SYSIN DD *
OPTION SKIPREC=1,EQUALS
INREC IFTHEN=(WHEN=GROUP,BEGIN=(1,4,CH,EQ,C'HEAD'),
PUSH=(3001:1,29))
SORT FIELDS=(3001,12,CH,A)
OUTREC IFTHEN=(WHEN=GROUP,BEGIN=(1,12,CH,EQ,C'HEAD CNTARXX'),
PUSH=(3030:ID=10)),
IFTHEN=(WHEN=GROUP,BEGIN=(1,12,CH,EQ,C'HEAD CNTBRXX '),
PUSH=(3030:ID=10))
OUTFIL FNAMES=OUT1,INCLUDE=(3001,12,CH,EQ,C'HEAD CNTARXX'),
REMOVECC,
BUILD=(1,3000),
TRAILER1=(3001,29,42:3030,10)
OUTFIL FNAMES=OUT2,INCLUDE=(3001,12,CH,EQ,C'HEAD CNTBRXX'),
REMOVECC,
BUILD=(1,3000),
TRAILER1=(3001,29,42:3030,10)
/*

by **akanchu** » Tue Aug 11, 2009 7:08 am

Hi Frank,
Thanks for your reply.
Do I have to have a specific version of SORT.
It goves me the following error

Select all

INREC IFTHEN=(WHEN=GROUP,BEGIN=(1,4,CH,EQ,C'HEAD'),
*
OUTREC IFTHEN=(WHEN=GROUP,BEGIN=(1,8,CH,EQ,C'...
*
WER268A INREC STATEMENT : SYNTAX ERROR
WER268A OUTREC STATEMENT : SYNTAX ERROR

We have version : SYNCSORT FOR Z/OS 1.3.1.0R

Am I missing something.
Thanks

by **dick scherrer** » Tue Aug 11, 2009 7:51 am

Hello,

We have version : SYNCSORT FOR Z/OS 1.3.1.0R

Am I missing something.

Yup. . . This part of the forum is for DFSORT topics.

I'll move the topic to the Syncsort part of the forum.

WHEN=GROUP is also supported by Syncsort, but your system support would need to install the latest upgrade.

IBM Mainframe Forum

Split file with key and its subsequent records

Split file with key and its subsequent records

Re: Split file with key and its subsequent records

Re: Split file with key and its subsequent records

Re: Split file with key and its subsequent records

Re: Split file with key and its subsequent records

Re: Split file with key and its subsequent records