RE: can table locking improve performance ??
- From: "Mercadante, Thomas F" <thomas.mercadante@xxxxxxxxxxxxxxxxx>
- To: "'oracle-l@xxxxxxxxxxxxx'" <oracle-l@xxxxxxxxxxxxx>
- Date: Tue, 8 Jun 2004 14:48:38 -0400
Babette,
My guess is that you will ether see no improvment at all, or it will get
worse.
Oracle will issue it's own locks - your placing a lock on a row may actually
slow it down. I can't see it getting any better. As you said, you will be
the only person on the database at the time - you will be contending with
anyone else trying to lock either the table or a row in the table.
I say leave it alone. Don't issue any locks at all - let Oracle take care
of them.
To speed up the process, you may need to re-engineer how you are doing it.
Are you performing a straight sqlldr load, CTAS, procedural
loop/insert/update?
Tom Mercadante
Oracle Certified Professional
-----Original Message-----
From: babette.turnerunderwood@xxxxxxxxxxxxx
[mailto:babette.turnerunderwood@xxxxxxxxxxxxx]
Sent: Tuesday, June 08, 2004 2:42 PM
To: oracle-l@xxxxxxxxxxxxx
Subject: can table locking improve performance ??
We have an initial load process that is very long running.
The update of 38M rows takes hours and hours.
There are two users on the entire system, us doing the
load and another user to look at some of the v$ views
during the load process.
Updates need to put exclusive row-level locks before
being able to update rows. If we do a table level lock
PRIOR to the update statement, what effect will it
have on performance. Intuitively, we think it should be
more efficient, but I don't know if Oracle will be more
efficient in checking locks when updating rows
or does the exact same source code apply, regardless
if there is a table level lock.
In other words, in this case, can a table level lock
improve performance ?
- thanks
Babette
Other related posts:
