View previous topic :: View next topic
|
Author |
Message |
HammerTime
New User
Joined: 26 Sep 2012 Posts: 14 Location: UK
|
|
|
|
Hi all, we use parallel processing in a number of our JCL's. Where the output is used against a specific table, we would choose the output 'split' based on the table partitioning ranges.
However, I have a series of jobs which run in parallel where we don't care so much about what goes in to each file, but what I would like to do is find a way of splitting the input into even sized output files.
For example:
Code: |
Input File: (12 recs)
----------
abc123
qwer7445
powe456
iopweg3456
poi345
heqw235
etio25
uewewe79789
we8323
opg732
poijjd4355
klre346
Desired outputs:
File1:
-------
abc123
qwer7445
powe456
File2:
-------
iopweg3456
poi345
heqw235
File3:
-------
etio25
uewewe79789
we8323
File4:
-------
opg732
poijjd4355
klre346
|
The number of records incoming daily will vary by a huge amount, so if I just use STARTREC and ENDREC it would mean on low volume days, I wouldn't be using all of the output files, and therefore not making the most of my parallel executions.
If there was a way I could dynamically spread the records across the outputs so they were all about even, it would allow the shortest possible execution time.
Is there any such functionality in DFSORT/ICETOOL? |
|
Back to top |
|
|
Dsingh29
Active User
Joined: 16 Dec 2008 Posts: 132 Location: IBM
|
|
Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011 Posts: 7309 Location: Inside the Matrix
|
|
|
|
Dave, Rohit too.
I'm going to ban just plain links to topics which are more than two years old.
HammerTime, certainly do as Dave says, and look at SPLIT and its cousins.
In attempting to do splits with STARTREC and ENDREC have you verified that you don't read the "front" of your file multiple times?
If SPLIT & Co don't do what you need, there is always a one-pass solution with code. |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10872 Location: italy
|
|
|
|
IMO
how to split a file should be added to the [in]famous category of
the topics no longer eligible for discussion |
|
Back to top |
|
|
|