On Tue, 25 Feb 2014 08:06:08 -0800, Sri h Kolusu wrote:
>Q . Most of the suggestions don't address Binyamin's concern with the
>limit on concatenation size.
>
>Paul,
>
>The Proposed solution by me indeed takes care of concatenation limits. I
>took the limit to be 254 datasets. The job I have shown will create 254
>job steps with each step having a concatenation limit of 254 datasets for
>SORTIN.
>
(IIRC, I've had other jobs fail at more like 1000 than 254.)
I'm sorry; I don't understand your SORT control statements. But still,
don't you wind up with several SORTOUTs where Binyamin wants one.
I suspect you have an easy answer for that, but does that address
Binyamin's other concerns:
On Mon, 24 Feb 2014 11:58:49 +0200, Binyamin Dissen wrote:
>
>... And I wish an all rexx solution.
>
>I ended up having step (1) combine them all into one big file and then doing a
>SORT. But it adds an extra pass across all of the data.
>
When one is proficient with a hammer, one treats everything as a nail.
So I wound up (in my imagination) with Rexx feeding a single SORTIN
via a POSIX pipe.
Other utilities on other OSes have a different solution:
561 $ uname -a
Linux Linux-Mac 3.2.0-59-generic-pae #90-Ubuntu SMP Tue Jan 7 23:07:06 UTC 2014
i686 i686 i386 GNU/Linux
562 $ man sort
...
--files0-from=F
read input from the files specified by NUL-terminated names in
file F; If F is - then read names from standard input
I suspect that Binyamin's Rexx code avoided the concatenation limit by using
a similar technique (as would mine).
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
