IBM Mainframe Forum

by **ranga_subham** » Sun Jan 25, 2009 7:10 pm

Hi,

I want to get the first 3 records for each key from the below dataset. I have pasted only portion of it (LRECL=190, RECFM=FB).

Select all

CENTRE1000
CENTRE1000
CENTRE1000
CENTRE1000
CENTRE1000
CENTRE9898
CENTRE9898
CENTRE9898
CENTRE9898
CENTR10980
CENTR10980
CENTR10980
CENTR10980
CENTR10980

Output:

Select all

CENTRE1000
CENTRE1000
CENTRE1000
CENTRE9898
CENTRE9898
CENTRE9898
CENTR10980
CENTR10980
CENTR10980

I tried using below code with SYNCTOOL but unsuccessfully. I thought it is not supported in v1.2.3.1R.

Select all

//TOOLIN DD *
SELECT FROM(IN1) TO(OUT1) ON(1,10,CH) FIRST(3)
/*

Please suggest.

Thanks.

by **Alissa Margulies** » Mon Jan 26, 2009 8:45 pm

ranga_subham,

If the data is already in sorted order, then you can use the job below:

Select all

//SORT1 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTOUT DD DSN=output.file,...
//SORTIN DD *
CENTRE1000
CENTRE1000
CENTRE1000
CENTRE1000
CENTRE1000
CENTRE9898
CENTRE9898
CENTRE9898
CENTRE9898
CENTR10980
CENTR10980
CENTR10980
CENTR10980
CENTR10980
//SYSIN DD *
INREC FIELDS=(1,10,11:SEQNUM,8,ZD,RESTART=(1,10))
SORT FIELDS=COPY
OUTFIL INCLUDE=(11,8,ZD,LE,3),OUTREC=(1,10)
/*

by **ranga_subham** » Tue Jan 27, 2009 5:09 pm

Wow.....thanks a lot Alissa.....nice alternative !

IBM Mainframe Forum

Get the first 3 records for each key

Get the first 3 records for each key

Re: Get the first 3 records for each key

Re: Get the first 3 records for each key