Re: Replacing a view with selects on it

From: Wolfgang Breitling <breitliw_at_centrexcc.com>
Date: Sun, 16 May 2004 11:40:21 -0600
Message-Id: <6.1.0.6.2.20040516113117.02b4e4f0@pop.centrexcc.com>

At 09:49 AM 5/16/2004, you wrote:

>1) start long select from a view
>
>2) create or replace the view while the session is running
>3) see what happen:)
>
> The reader session did "survived" (I guess with the cached version of
> the view).

Unless my understanding is utterly wrong, the view definition is only required during the parsing of the query. Once it starts the execute/fetch phase the plan is bound to the physical objects and the view definition can change.
I am confident that Oracle would prevent view changes until the query finishes if it was needed beyond the parse.

>But it is no prove!
>
>Are there situation, where this approach can cause problems? Any
>experience with this topic on the list?

Based on my understanding and assessment, I do not foresee any problems. I have not done it on a massive or regular basis, but I have never had a problem altering a view definition in a live system, except when not using replace and permissions get lost temporarily.

>Jaromir D.B. Nemec

regards

Wolfgang Breitling
Centrex Consulting Corporation
www.centrexcc.com

Please see the official ORACLE-L FAQ: http://www.orafaq.com

To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line.

--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Received on Sun May 16 2004 - 12:37:51 CDT