MVSFORUMS.com

[PaulPeplinski]

I have two files with a one to many relationship where JOINKEYS inner join seems to be the ticket - except that two one-byte fields need to be set with constant values indicated with the "?" below. In this case I need "22" at location 53 of the output record and 2's are not in either record.

I imagine I could do the joinkeys first and then OUTREC them in (what is likely) a second pass. If that's the case I am probably defeating the purpose of trying to improve the performance of the current COBOL implementation (both CPU and runtime).

In this case each file is over 80 million records.

I assume this function needs the SORT FIELDS statement.

[kolusu]

[dz]

You can do OUTREC in the same step, after the REFORMAT.

[kolusu]

[amargulies]

[kolusu]

[amargulies]

Kolusu,

SyncSort for z/OS 1.2 supports JOIN. JOIN processing logically replaces the reading of the SORTIN datasets. Other Phase 1 features, like INREC, INCLUDE, OMIT, and STOPAFT still can be used. While INCLUDE/OMIT can be specified on the JOINKEYS statement as well as a traditional INCLUDE/OMIT statement, INREC can modify a record created by the JOIN processing.

Here is the example you requested:

[kolusu]

amargulies,

I am aware of the INREC statement , but my question was to use seperate inrec for each SORTJNF DD.

something like this

[amargulies]

[PaulPeplinski]

Thank you. The job is set to go with the following control

[Phantom]

Thats nothing....Sort products gives you a warning when you try to increase the LRECL of the input file using INREC - bcas it has to process a much larger file than actual. That message is just an indication to use OUTREC or OUTFIL OUTREC to reformat records (whenever you can).

Thanks,
Phantom