Re: driving_site hint in Materialized View definition
From: joel garry <joel-garry_at_home.com>
Date: Wed, 1 Jul 2009 09:12:07 -0700 (PDT)
Message-ID: <f00f413d-7380-44bf-bbc2-ae3f626e88dd_at_t21g2000yqi.googlegroups.com>
On Jul 1, 12:28 am, "Jonathan Lewis" <jonat..._at_jlcomp.demon.co.uk> wrote:
> "Yong Huang" <yong..._at_yahoo.com> wrote in message
>
> news:06ac8013-c447-45b7-bc73-8a8cd452329b_at_l32g2000vba.googlegroups.com...
>
>
>
> > On Jun 27, 9:11 am, "Jonathan Lewis" <jonat..._at_jlcomp.demon.co.uk>
> > wrote:
>
> >> The driving_site() hint is not legal in "create as select"
> >> or "insert as select". The select statement has to operate
> >> at the database where the create or insert takes place.
>
> >> I think if you search metalink, you'll find a note that makes
> >> this point - but I don't have a reference to the doc ID.
>
> > Metalink Note:825677.1 says in
>
> > insert /* DRIVING_SITE(remote_example_data) */ into example_data
> > select /* DRIVING_SITE(remote_example_data) */ remote_example_data.*
> > from ...
>
> > the first driving_site hint, the one after insert, is ignored. That
> > must be the note you're referring to?
>
> > Yong Huang
>
> I should have remembered I had a note on my blog about
> distributed DML and the irrelevant of the driving_site() hint
> for DML. It references the metalink note - and reminds me
> that it was Joel Goodman who pointed it out to me.
>
> http://jonathanlewis.wordpress.com/2008/12/05/distributed-dml/
>
Date: Wed, 1 Jul 2009 09:12:07 -0700 (PDT)
Message-ID: <f00f413d-7380-44bf-bbc2-ae3f626e88dd_at_t21g2000yqi.googlegroups.com>
On Jul 1, 12:28 am, "Jonathan Lewis" <jonat..._at_jlcomp.demon.co.uk> wrote:
> "Yong Huang" <yong..._at_yahoo.com> wrote in message
>
> news:06ac8013-c447-45b7-bc73-8a8cd452329b_at_l32g2000vba.googlegroups.com...
>
>
>
> > On Jun 27, 9:11 am, "Jonathan Lewis" <jonat..._at_jlcomp.demon.co.uk>
> > wrote:
>
> >> The driving_site() hint is not legal in "create as select"
> >> or "insert as select". The select statement has to operate
> >> at the database where the create or insert takes place.
>
> >> I think if you search metalink, you'll find a note that makes
> >> this point - but I don't have a reference to the doc ID.
>
> > Metalink Note:825677.1 says in
>
> > insert /* DRIVING_SITE(remote_example_data) */ into example_data
> > select /* DRIVING_SITE(remote_example_data) */ remote_example_data.*
> > from ...
>
> > the first driving_site hint, the one after insert, is ignored. That
> > must be the note you're referring to?
>
> > Yong Huang
>
> I should have remembered I had a note on my blog about
> distributed DML and the irrelevant of the driving_site() hint
> for DML. It references the metalink note - and reminds me
> that it was Joel Goodman who pointed it out to me.
>
> http://jonathanlewis.wordpress.com/2008/12/05/distributed-dml/
>
So does the database machine moot this or make it more important?
jg
-- _at_home.com is bogus. WTF Bob? http://www.weblog.keepthejointrunning.com/Received on Wed Jul 01 2009 - 11:12:07 CDT