IBM Mainframe Forum

A couple more things. You're not actually doing any "incrementing". Did you go with the idea of maintaining a run-sequence control file separately?

Also, if you are able to identify the header, then you'd not need the sequence-number generation for each record.

Hi Billy, yea for some reasons business has asked us to maintain 15 digit seq.

Also for SYMNAMES I created a flat file with LRECL=80, RECFM=FB instead of using a temp data set.

Also, with the tem dataset name changed to &&temp* the job worked fine

Up to 999,999,999,999,999 records? of 536 bytes each?

OK, it is really, really, going to help if you can identify the header without using the sequence-number approach, if you have a file even remotely approaching that size.

It would frankly be "nuts" to pass through some huge amount of data, just to update the sequence number for such a file.

I think you need to get some clarification. Find out the expected initial size of the file, dummy up some records, find how long it takes, take file growth into account and do some estimates for one year, two years, whatever, then go back to the requirement-giver and see if they are happy.

As the sort product in question is SYNCSORT (WER messages)I have moved the topic to that forum.

Hi,

I am using the same JCL for incrementing sequence numbers. I am facing the same issue as mentioned by you:

This job is going thru fine but is not appending the sequence number in File 2.
If you could kindly assist me where I am going wrong.

Can you please tell, how this was corrected?