>> We run sort splits on large tape datasets, with over a hundred million
>> records. They get split into 3 separate datasets, depending on the data.
These jobs do not use FILSZ.
Eric,
Does the job SORT the data before split , or is it just splitting with a COPY
operation to separate datasets ? As long the tape datasets have a standard
label and Tape processing software(RMM/CA-1/ICETPEX) is providing the details
of the dataset, you don't need to code FILSZ parameter. If your job has
SORTDIAG DD DUMMY coded, then you can check the ICE905I message.
for example.
ICE905I 1 I : RF=144,LR=340,BLK=17680,BCT=514473
I = INPUT (SORTIN OR SORTINXX)
RF=144 RECFM 128 = X'80' F - FIXED +
16 = X'10' B - BLOCKED
LR=340 IS THE LRECL OF THE FILE
BLK=17680 IS THE BLOCKSIZE OF THE DATASET
BCT=514473 BLOCK COUNT (WILL BE 0 FOR COPY, MERGE OR SORTOUT)
Thanks,
Kolusu
DFSORT Development
IBM Corporation
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
