Re: Library Cache question

From: Cary Millsap <cary.millsap_at_method-r.com>
Date: Fri, 6 Feb 2009 14:07:26 -0600
Message-ID: <3a2a84fc0902061207p2386e3c0h7de76429f5e15924_at_mail.gmail.com>

...And I *do* know from testing I conducted a couple of years ago that using SIMILAR does marginally diminish the parse performance of well-written application code relative to using FORCE.

Cary

On Fri, Feb 6, 2009 at 2:03 PM, Jared Still <jkstill_at_gmail.com> wrote:

> On Fri, Feb 6, 2009 at 8:06 AM, Martin Brown <martinfbrown_at_hotmail.com>wrote:
>
>>
>> Thanks for the reply, Gentlemen. We use "FORCE". I guess my next step is
>> to test "SIMILAR" at some point and compare the results.
>>
>
> Ah, that brings up more questions.
>
> Caveat: I have not done testing of 'FORCE' so some of my questions may be
> naive.
>
> How much time is being recorded against library cache misses?
>
> Are new plans actually being generated for these statements?
>
> The reason I ask is that I don't know at what point Oracle determines
> that an existing query plan may be used for a SQL statement that is
> identical to one already seen, save for the use of literals in the
> predicate.
>
> If Oracle is recording a cache miss, and then the CBO determines that
> a new plan is not needed, the database may not actually be charging
> much time against the cache misses.
>
> That statement may be ignorance on my part, but I just don't have time
> right now to see how that works.
>
> Whereas you have more incentive to do so. :)
>
> Jared Still
> Certifiable Oracle DBA and Part Time Perl Evangelist
>
>

--
http://www.freelists.org/webpage/oracle-l

Received on Fri Feb 06 2009 - 14:07:26 CST