Commit Frequency
Moderators: chulett, rschirm, roy
Commit Frequency
How to handle commit frequency in DataStage Parallel Jobs and where this property is to set in job. :?:
-
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
-
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
-
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
Commit Interval
If you want to do commit on Updates in DB2:
InDB2 Enterprise stage we can specify when to commit by using two parameters:
1) Time Commit Interval (2 seconds by default)
2) Row Commit Interval (2,000 by default. We can keep its value only in multiples of Array Size(default 2,000))
But what we found is that Row Commit Interval parameter is of no use as whatever value you give to Array Size parameter, commit happens on that. So let's say if Row Commit Interval=10,000 and Array Size=5,000 and the job aborts after updating 9,000 records instead of rolling back to 0 it rolls back to 5,000 (Array Size).
So one can keep Array Size and Time Commit Interval parameters. Commit will happen for lesser of the two parameter values.
InDB2 Enterprise stage we can specify when to commit by using two parameters:
1) Time Commit Interval (2 seconds by default)
2) Row Commit Interval (2,000 by default. We can keep its value only in multiples of Array Size(default 2,000))
But what we found is that Row Commit Interval parameter is of no use as whatever value you give to Array Size parameter, commit happens on that. So let's say if Row Commit Interval=10,000 and Array Size=5,000 and the job aborts after updating 9,000 records instead of rolling back to 0 it rolls back to 5,000 (Array Size).
So one can keep Array Size and Time Commit Interval parameters. Commit will happen for lesser of the two parameter values.