On Oct 4, 2007, at 2:03 PM, Steve Comstock wrote:
Amlan Prasad wrote:
Hi,
Is it possible in the JCL to avoid coding LRECL for datasets which
created new. If it can be automatically understood based on the
COBOL program creating then any change to the code will not
require JCL changes. We have frequent file length changes and
often some JCL changes are missed and ABEND is faced. I am aware
it is not mandatory to code for SORT. It will pick the LRECL for
SORTOUT based on SORTIN.
Thanks,
Amlan
It has always been possible and always recommended _not_
to code LRECL in JCL for COBOL datasets. COBOL _always_
figures out the LRECL for you. And you should code
BLOCK CONTAINS 0 RECORDS except, possibly, for print
files, where the default of BLOCK CONTAINS 1 RECORDS
works fine (because spool is blocked by JES in its
own manner anyway).
Steve,
Well its a mixed bag. I have seen small reports (1 page) become 100's
(or more) pages. While its true that it doesn't make much different
for a JES dataset, its different if its changed to tape or disk.
(real or virtual) I would still recommend that block contains 0
records for the vast majority of FD's.
Ed
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
