This question rolled into my In Box today: In the case of using the LIMIT clause of BULK COLLECT, how do we decide what value to use for the limit? First I give the quick answer, then I provide support for that answer Quick Answer Start with 100. That's the default (and only) setting for cursor FOR loop optimizations. It offers a sweet spot of improved performance over row-by-row and not-too-much PGA memory consumption. Test to see if that's fast enough (likely will be for many cases). If not, try higher values until you reach the performance level you need - and you are not consuming too much PGA memory. Don't hard-code the limit value: make it a parameter to your subprogram or a constant in a package specification. Don't put anything in the collection you don't need. [from Giulio Dottorini] Remember: each session that runs this code will use that amount of memory. Background When you use BULK COLLECT, you retrieve more than row with each fetch, ...
The commit was within an autonomous transaction.
ReplyDeleteAutonomous Transaction!
ReplyDeletecommit write nowait/batch?
ReplyDeleteAutonomous Transaction :allows you to leave the context of the calling transaction, perform an independent transaction, and return to the calling transaction without affecting it's state.Hence the uncommitted changes in the users session are still significant.
ReplyDeleteAutonomous transaction! Love the brain teaser!!
ReplyDeleteJust few untested random use-cases in a hurry. Feel free to correct.
ReplyDelete1. Dirty data is in bulk collected collections. Processing commits in loop.
2. View with instead of trigger. Trigger only uses partial data to update.
3. Part of dirty data is written to a external table on really slow IO device.
4. Distributed transaction. Part of dirty data has to be updated on some other DB.
5. Call to external service from API with part of dirty data.
6. Trigger on table allows selective data to update?
It all depends where in the process chain the commit was executed, if there were any dml statements after it, and if any rollback (to savepoint) was executed...
ReplyDeleteAnd here are some thoughts offered up on LinkedIn:
ReplyDelete* Is the commit statement outside the procedure? It so, it may be out of scope. Is there a rollback somewhere in the procedure or function?
* The procedure having COMMIT might have executed with autonomously. So the user changes outside that procedure are still in process and not yet committed.
And now my answer: definitely, the answer in my mind is that the COMMIT statement was executed within an autonomous transaction subprogram.
ReplyDeleteIf you include
PRAGMA AUTONOMOUS_TRANSACTION;
in the declaration section of your procedure or function, then a COMMIT in that subprogram will commit only those changes made in the scope of that subprogram.
Other outstanding changes in my session will NOT be committed.
Now as to your comments:
@stanley, I'd love to hear more explanation about some of your items, as they are outside my area of expertise. I am not sure if a distributed xaction applies here, since I reference the "user's session".
@john, certainly any DML statements executed after the commit would be uncommitted. ROLLBACK TO before the commit would remove outstanding changes. After the commit? Well, the teaser has to do with state of session right after the commit (implied: before other actions take place).
Thanks for participating!