Re: index skip scan

From: Robert Klemme <shortcutter_at_googlemail.com>
Date: Tue, 06 Jan 2009 17:19:32 +0100
Message-ID: <6shekkF61vo3U1_at_mid.individual.net>

On 06.01.2009 16:44, jernigam_at_nospam.kochind.com wrote:
> Yes it is a DATE column. I get the same plan with to_date.

And your statistics are of course current and accurate, aren't they?

> I know it doesn't use bind variables. This is a simplified version
> of multiple queries that are in a packaged application. I can't
> change the queries in the app.

So, are you seeing the bad plan with the simplified query or with the real query (or both)? I am not sure what to think of this as there seem to be a lot of details missing.

> I am trying to understand what impacts the cost of index skip scans.
> The 10053 trace shows the cost of the index skip scan as being
> lower.

You should also be able to determine _why_ it is lower from the trace. It should at least give you an idea.

> I hate to delete the index used by the skip scan, because it is
> probably needed by some other queries.

Did you try an SQL Profile?

Cheers

        robert

-- 
remember.guy do |as, often| as.you_can - without end

Received on Tue Jan 06 2009 - 10:19:32 CST