IBM Mainframe Forum

Hi,

The JOINKEYS is causing the output files in the sorted order though SORT FIELDS=COPY is mentioned. I don't want the SORT to happen but copy the records as they are from input to output.

How to achieve it?

TIA.

Hello,

Is there some reason you did not post the jcl and sort control statements (use the Code tag)?

I don't want the SORT to happen but copy the recoreds as they are from input to output.

Confusing

JOINKEYS implies the data will be insequence. . .

Dick, my JCL and SORT card everything is working and am getting output but the only problem is data is appearing in sorted order. I don't want this to happen.

Actually, I am picking unique values from a file and trying to extract those values from some other huge file.

Thx.

Hello,

And the jcl/sort control has still not been posted. . .

To repeat - JOINKEYS must have data in sorted order. . . It does not provide a random file search.

my JCL and SORT card everything is working and am getting output but the only problem is data is appearing in sorted order

Which means it is not really working. . .

dick scherrer wrote:To repeat - JOINKEYS must have data in sorted order. . . It does not provide a random file search.

Hi Dick,

May be I misunderstood your comment here. But JOINKEYS does n't need the input data to be in sorted order of the key. But the JOIN operation does sort both the input files on the key mentioned in the JOINKEYS statement before doing a file-match.

dick scherrer wrote:Actually, I am picking unique values from a file and trying to extract those values from some other huge file

ranga_subham,

Is your 'huge file' sorted on any field? Please provide the details requested here.

Hi Arun,

Thanks for the correction. Dain Brammage. . .

I have no idea why, but i was thinking of a merge operation

Keep me in line


Thanks again,
d

Hi Dick,

You're welcome.

If the original input files are "random" with no particular sequence, and the requirement is to maintain this random order in the output, you can pre-process the datasets to include a sequence number. Sort on that SEQNUM after JOIN processing, and then use OUTREC to strip off the SEQNUM.