IBM Mainframe Forum

by **helen2000** » Tue Aug 11, 2009 9:53 am

Hi All,

I have a trouble during sorting. I can sort all records except the header. because my header record didn't have
sort key. does anybody have good idea for that first copy the header record and then sorting the others records.
my sort production is Syncsort for Z/os.

thanks.

by **nayanpatra** » Tue Aug 11, 2009 12:30 pm

STEP010: First break the file into two files using STOPAFT = 1 for the header and SKIPREC = 1 to capture the records without header. Remember to use SORT FIELDS = COPY.
STEP020: Now sort the second file containing only the records acc. to your requirements.
STEP030: Next combine the two files into a single file. For this you can use REPRO and give the two files as the SORTIN DD statement. In the control card use SORT FIELDS = COPY.

by **arcvns** » Tue Aug 11, 2009 6:32 pm

Hello,

Here's a one-step SyncSort job which does exactly what you asked. This is just a sample, you can easily modify it as per your requirement.
Let me know if you come across any issues.

Select all

//STEP1 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTIN DD *
HEADER
KEY-A
KEY-C
KEY-D
KEY-B
//SORTOUT DD SYSOUT=*
//SYSIN DD *
INREC IFTHEN=(WHEN=INIT,OVERLAY=(81:SEQNUM,8,ZD,C'1')),
IFTHEN=(WHEN=(81,8,ZD,EQ,1),OVERLAY=(89:C'0'))
SORT FIELDS=(89,1,CH,A,1,5,CH,A)
OUTFIL BUILD=(1,80)

SORTOUT

Select all

HEADER
KEY-A
KEY-B
KEY-C
KEY-D

by **ranga_subham** » Tue Aug 11, 2009 9:18 pm

Arun, what a cool trick this is.......

Very nice.

by **helen2000** » Tue Aug 11, 2009 11:49 pm

thanks all of you.
it works well right now. but there is
a little bit problem. my input record is VB. althougn I
got the output dataset. but the header record is
extended to 2004. Could you tell me how to keep the same
size as before sorting for the header record. thanks.

Select all

026900 //SYSIN DD *
026910 INREC IFTHEN=(WHEN=(6,5,ZD,EQ,12251),
026920 OVERLAY=(249:C'0',2004:C'0'))
027000 SORT FIELDS=(40,10,CH,A,
027100 11,8,CH,A,
027200 249,2,BI,A)
027210 //* SKIPREC=1
027300 /*

by **dick scherrer** » Wed Aug 12, 2009 12:30 am

Hello,

You might consider placing the seqnum at the beginning of the records instead of the end. . .

by **helen2000** » Wed Aug 12, 2009 1:30 am

thanks, dick

maybe I didn't describe the problem clear. the input dataset have 3 sort key(offset is 40,11 and 249).
all input record have the 3 sort key except header(just 2 sort key, because header's size is 166).
when I run the JCL, got the error message:" control field exceed size". that's why I have to extend
the header record size before sorting.

by **Alissa Margulies** » Wed Aug 12, 2009 1:32 am

Helen2000,

Do you want the header record to remain as the first record, or to be sorted along with all the data records?

by **arcvns** » Wed Aug 12, 2009 10:05 am

helen2000,

A few more questions apart from what Alissa has asked.

Does "(6,5,ZD,EQ,12251)" identify your header record?
What is your input LRECL?

by **nayanpatra** » Thu Aug 13, 2009 10:16 am

You could convert from variable format to fixed format.

IBM Mainframe Forum

how to copy header record during sorting.

how to copy header record during sorting.

Re: how to copy header record during sorting.

Re: how to copy header record during sorting.

Re: how to copy header record during sorting.

Re: how to copy header record during sorting.

Re: how to copy header record during sorting.

Re: how to copy header record during sorting.

Re: how to copy header record during sorting.

Re: how to copy header record during sorting.

Re: how to copy header record during sorting.