IBM Mainframe Forum

Hi Frank, Thanks a lot

It is working as expected... Thanks once again. I have a doubt about CPU. Will it consume more CPU if we have 1 lakh records on each input file as we are using two temp files in a single icetool step. Please advice..

Consume more CPU time compared to what?

1 lakh is only 100000 records which is a rather small file as far as DFSORT is concerned. The CPU time consumed is related to the number of passes over the data, but the job won't consume much CPU time for that small a file.

Hi Frank,

Thanks a lot. It is working as expected. Thanks once again.
I need small clarification regarding CPU usage. Will this icetool step consume more CPU if each of the four input files has approximately 1 lakh records as we are using two temparory datasets in single icetool step.

Thank you.

Hi Frank,

Sorry for posting same question again. Please ignore it.

Hi,
Is there any utility avaiable for doing this without creating any kind of intermediate datasets?

gokulNmf wrote:Hi,
Is there any utility avaiable for doing this without creating any kind of intermediate datasets?

gokulNmf,

If all 4 files are of the same LRECL and you have a unique indicator in all of the four files that identifies at to which file the record belongs to , then DFSORT with WHEN=GROUP will get you the desired results without using any intermediate files.

Thanks skolusu !