IBM Mainframe Forum

by **ibmmf4u** » Fri May 18, 2012 11:48 pm

Hi Everyone,

My requirement goes this way, I would like to find out the maximum value of a particular group and write it to an output file. Mentioned below is an example.

Input file:-

Select all

TEST1 0010
TEST1 0020
TEST2 0030
TEST2 0040
TEST3 0050
TEST3 0060
TEST3 0070
TEST4 0080
TEST4 0060
TEST5 0090
TEST5 0100

Output file:-

Select all

TEST1 0020
TEST2 0040
TEST3 0070
TEST4 0080
TEST5 0100

Can someone provide me the sort card in achieving the above

Thanks in advance!!

by **ibmmf4u** » Sat May 19, 2012 11:38 am

Hi All,

I can able to achieve the above with synctool. Pasted below is the code that's working fine.

Select all

//STEP010 EXEC PGM=SYNCTOOL
//TOOLMSG DD SYSOUT=*
//SSMSG DD SYSOUT=*
//INPUT DD *
TEST1 0010
TEST1 0020
TEST2 0030
TEST2 0040
TEST3 0050
TEST3 0060
TEST3 0070
TEST4 0080
TEST4 0060
TEST5 0090
TEST5 0100
//OUTPUT DD SYSOUT=*
//TOOLIN DD *
SELECT FROM(INPUT) TO(OUTPUT) ON(1,5,CH) FIRST USING(CTL1)
/*
//CTL1CNTL DD *
SORT FIELDS=(1,5,CH,A,10,4,ZD,D)
/*

Its giving me the exact output i.e. required.

However earlier, i am able to sort the records in the order i.e.required as per the output with the below sort card

Select all

SORT FIELDS=(1,5,CH,A,10,4,ZD,D)

But can't able to eliminate the duplicates as per the required output. Pasted below is the output file that was generated with the help of the above sort card.

Select all

TEST1 0020
TEST1 0010
TEST2 0040
TEST2 0030
TEST3 0070
TEST3 0060
TEST3 0050
TEST4 0080
TEST4 0060
TEST5 0100
TEST5 0090

Can someone provide me the sort card on how to eliminate the duplicates from the above the file and get the required output.

Thanks in advance!!!

by **dick scherrer** » Sat May 19, 2012 8:04 pm

Good to hear it is working - thank you for letting us know

d

by **ibmmf4u** » Sat May 19, 2012 9:52 pm

Thank you Dick!!

Can you please let me know ho w to eliminate the duplicates from the below intermediate file .

Select all

TEST1 0020
TEST1 0010
TEST2 0040
TEST2 0030
TEST3 0070
TEST3 0060
TEST3 0050
TEST4 0080
TEST4 0060
TEST5 0100
TEST5 0090

Thanks in advance

by **ibmmf4u** » Sun May 20, 2012 2:37 pm

Hi All,

I can able to achieve it with help of sort. pasted below is the code that's working fine.

Select all

//STEP010 EXEC PGM=SORT
//SORTIN DD *
TEST1 0010
TEST1 0020
TEST2 0030
TEST2 0040
TEST3 0050
TEST3 0060
TEST3 0070
TEST4 0080
TEST4 0060
TEST5 0090
TEST5 0100
//SORTOUT DD DSN=TEST.SORT.RES,DISP=SHR
//SYSOUT DD SYSOUT=*
//SYSIN DD *
SORT FIELDS=(5,5,CH,A,14,4,ZD,D)
/*
//STEP020 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTIN DD DSN=TEST.SORT.RES,DISP=SHR
//SORTOUT DD SYSOUT=*
//SYSIN DD *
SORT FIELDS=(5,5,CH,A)
SUM FIELDS=NONE
/*

by **BillyBoyo** » Sun May 20, 2012 10:19 pm

OK, you might also want to see if your release supports WHEN=GROUP.

If, so, you should be OK with one step.

Do the sort (from your first step, with the A/D depending on whether you want to retain lowest/highest amount), then with OUTREC (which is processed after SORT/SUM) use WHEN=GROUP to set a sequence number.

With OUTFIL, INCLUDE only those records with sequence number of one, then return record to desired length (with BUILD, for instance).

by **dick scherrer** » Sun May 20, 2012 10:36 pm

Hello,

I didn't completely comprehend what i replied to last time. My bad :oops:

Do try what BB has suggested and try to have only one step. It there only 100 records, 2 steps might not be so bad, but if there are 100million records. . .

by **ibmmf4u** » Sun May 20, 2012 11:22 pm

Thanks Bill & Dick.

I shall try the same and let you know the outcome.

by **bodatrinadh** » Mon May 21, 2012 3:21 pm

Hello ibmmf4u,

If you are using SYNCSORT FOR Z/OS 1.4.0.1R version, you can try below code.

Select all

//STEP1 EXEC PGM=SORT
//SYSPRINT DD SYSOUT=*
//SYSOUT DD SYSOUT=*
//SORTIN DD *
TEST1 0020
TEST2 0030
TEST2 0040
TEST3 0050
TEST3 0060
TEST3 0070
TEST4 0080
TEST4 0060
TEST5 0090
TEST5 0100
//SORTOUT DD SYSOUT=*
//SYSIN DD *
SORT FIELDS=(1,5,CH,A)
DUPKEYS MAX=(10,4,ZD)

Output:-

Select all

TEST1 0020
TEST2 0040
TEST3 0070
TEST4 0080
TEST5 0100

Thanks
-3nadh

by **ibmmf4u** » Mon May 21, 2012 9:49 pm

Hi 3nadh,

Thanks a lot. we are currently using syncsort 1.3.2.2R. Its working fine here.

Regards,

IBM Mainframe Forum

Maxvalue of a Group

Maxvalue of a Group

Re: Maxvalue of a Group

Re: Maxvalue of a Group

Re: Maxvalue of a Group

Re: Maxvalue of a Group

Re: Maxvalue of a Group

Re: Maxvalue of a Group

Re: Maxvalue of a Group

Re: Maxvalue of a Group

Re: Maxvalue of a Group