Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: fine grain access control + partitioned table

Re: fine grain access control + partitioned table

From: Jaap W. van Dijk <j.w.vandijk_at_hetnet.nl>
Date: Mon, 21 Jan 2002 20:50:22 GMT
Message-ID: <3c4c7d0a.819108@news.freeler.nl>


Your posting is rather unclear to me.

What do you mean by 'placing a policy'?
How does that 'return a predicate'?
Is this relevant to your concern?
What does your explain plan look like ( so as to check that your interpretation is the same as that of other people)? How does this relate to truncating a partition?

If you query a partition, Oracle will not access other partitions, so I'm very curious to what your explain plan looks like. If you truncate a partition, Oracle will truncate the partition and nothing more.

Jaap.

On Mon, 21 Jan 2002 22:25:39 +1100, "G.Ong" <onggs_at_acslink.aone.net.au> wrote:

>Hi,
>
>Our system is 8.1.6.3 on Solaris 2.7.
>We came across this pecularity:
>
>A policy is placed on a partitioned table; the predicate returned is 1=1 or
>2=1 (to simplify discussion).
>If I perform this query:
>
> select count(*) from table_x partition (part_1) ;
>
>the explain plan shows that all partitions are scanned. The execution took a
>long time. The concern is that if I truncate a partition, it may truncate
>the whole table!!
>
>Have anybody come across this ?
>
>TIA,
>Ghee
>
>
Received on Mon Jan 21 2002 - 14:50:22 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US