IBM Mainframe Forum

by **aaa** » Wed May 11, 2011 8:12 pm

I'm using JOINKEYS to keep all records from a dataset (F1) that have a paired record in (F2). I'd like to remove duplicated keys from F2 to avoid the generation of duplicated F1 records in the output file, and since F2 has been previously sorted, I have specified the SORTED option for F2 and I'm using the SUM FIELDS=NONE statement to remove duplicates before the JOIN phase.

MERGE is selected by DFSORT from the F2 subtask, but it seems to ignore the SUM FIELDS=NONE statement. If I don't use the SORTED option it behaves as expected, but I'll like to avoid sorting an already sorted dataset. Is there a way to remove duplicates from F2 without using another job step?

Job step

Select all

//STEP010 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//MASTER DD *
1AAAAAAAAA
5BBBBBBBBB
3CCCCCCCCC
4DDDDDDDDD
//PULL DD *
3
5
5
5
7
//SORTOUT DD SYSOUT=*
//DFSPARM DD *
* CONTROL STATEMENTS FOR JOINKEYS APPLICATION
JOINKEYS F1=MASTER,FIELDS=(1,1,A)
JOINKEYS F2=PULL,FIELDS=(1,1,A),SORTED
REFORMAT FIELDS=(F1:1,10)
* CONTROL STATEMENT FOR MAIN TASK
OPTION COPY
//JNF2CNTL DD *
* CONTROL STATEMENT FOR SUBTASK2 (F2)
SUM FIELDS=NONE
/*
//*

Actual OUTPUT

Select all

3CCCCCCCCC
5BBBBBBBBB
5BBBBBBBBB
5BBBBBBBBB

Expected OUTPUT

Select all

3CCCCCCCCC
5BBBBBBBBB

by **aaa** » Wed May 11, 2011 8:25 pm

I just read this in the DFSORT manual:

If you use the SORTED operand, statements and options only available for a sort application, such as SUM, will be ignored for the subtask that copies the input file.

It looks like it's not possible to remove duplicated without sorting the dataset. But since it's using a MERGE application, shouldn't the SUM statement be supported?

by **skolusu** » Wed May 11, 2011 9:21 pm

aaa,

When you use SORTED operand , DFSORT Simply copies the records from the file and SUM statement does not apply for a COPY statement as there is no key to compare and remove the duplicates. Use the following DFSORT control cards which will give you the desired results

Select all

//DFSPARM DD *
* CONTROL STATEMENTS FOR JOINKEYS APPLICATION
JOINKEYS F1=MASTER,FIELDS=(1,1,A)
JOINKEYS F2=PULL,FIELDS=(1,1,A),SORTED
REFORMAT FIELDS=(F1:1,10,F2:3,5)
* CONTROL STATEMENT FOR MAIN TASK
OPTION COPY
OUTFIL BUILD=(1,10),INCLUDE=(11,5,ZD,EQ,1)
//JNF2CNTL DD *
* CONTROL STATEMENT FOR SUBTASK2 (F2)
INREC OVERLAY=(3:SEQNUM,5,ZD,RESTART=(1,1))
/*
//*

IBM Mainframe Forum

Removing duplicates from sorted dataset when using JOINKEYS

Removing duplicates from sorted dataset when using JOINKEYS

Re: Removing duplicates from sorted dataset when using JOINK

Re: Removing duplicates from sorted dataset when using JOINK