Re: Temporary table use and the redo log
From: Mladen Gogala <gogala.mladen_at_bogus.email.invalid>
Date: Thu, 10 Sep 2009 11:17:17 +0000 (UTC)
Message-ID: <h8an7t$lsi$2_at_solani.org>
Na Wed, 09 Sep 2009 13:20:18 -0700, Randolf Geist napisao:
Date: Thu, 10 Sep 2009 11:17:17 +0000 (UTC)
Message-ID: <h8an7t$lsi$2_at_solani.org>
Na Wed, 09 Sep 2009 13:20:18 -0700, Randolf Geist napisao:
> Please check above mentioned point: You need to verify that the inserts
> are actually direct-path inserts in your "actual problem stored
> procedure". You can do so by using above technique of querying/
> accessing the table after inserting before committing. If this *doesn't*
> fail with an ORA-12838, then you've performed a conventional insert. The
> "APPEND" hint will be ignored if any of the restrictions apply for
> direct-path inserts.
Interesting method for establishing success or failure of the direct-path insert. I usually trace the session and see if there are direct file writes in the wait events.
-- http://mgogala.freehostia.comReceived on Thu Sep 10 2009 - 06:17:17 CDT