Author: TS70363
Posted: Tue Sep 13, 2016 3:26 am (GMT 5.5)
Thanks all for your valuable suggestions.
We use MSUs and it is a minimum of 5000 records per MSU.
I made the change and moved the SQLs within one of the SP to the program. I then, did a runstats and rebind and it proved to be fruitful. I got a run rate of 7 CPU minutes for 200000 records.
But the questions still remain unanswered. Why does the stored proc has poor performance in batch compared to embedded SQLs?
_________________
TS70363
Posted: Tue Sep 13, 2016 3:26 am (GMT 5.5)
Thanks all for your valuable suggestions.
| Quote: |
| What do other batch DB2 programs AT YOUR SITE use for CPU time per record? |
| Quote: |
| You don't have to make the change for 12 becaue you can find one SP which you think is a candiDate for poor performance so save the calling parameter somewhere and then put the same inout parms and SQL in a cobol db2 sug program and mock the same situation. I don't think this would take more time even if it takes some time that's well justified |
But the questions still remain unanswered. Why does the stored proc has poor performance in batch compared to embedded SQLs?
_________________
TS70363