MVSFORUMS.com

[mahesh_chv]

Hi ,

I have a requirement where in the our mainframe file it sent to java for further processing. whenever the file size is big , its failing on the java side. Whenever this happens , we used to split our mainframe file by using sync sort and send the files for further processing.

We would like to automate this process and in that process I would like to know how can we create the datasets dynamically ?

for example,

consider a scenario where in my Mainframe file is having 52,300 records.
I would like to split this file into smaller ones containing 10000 records each. The mainframe job should create 6 datasets ( 5 datasets with 10000 records and the 6 th one with 2300 records ) . Right now we manually check the file for the total count of the records and decide the number of datasets we need to create and go for a sync sort split .

tomorrow if the Mainframe files contains 45000 records, the jcl should be intelligent enough to split it into 5 files ( 4 files with 10k records and 5 th with 5k records ) .

Can I acheive this without manually checking the total records in the file and deciding the number of files it has to split ?


Thanks,
Mahesh

[kolusu]

mahesh_chv,

Check the smart DFSORT trick "Split a file to n output files dynamically" here

http://www.ibm.com/systems/support/storage/software/sort/mvs/tricks/index.html

Hope this helps...

Cheers
_________________
Kolusu
www.linkedin.com/in/kolusu

[mahesh_chv]

Thank you kolusu .....
But we use Sync Sort in our shop

[Dibakar]

Look for SPLIT1R in the link Kolusu provided. The trick should work for Syncsort too.