Re: OT : 9i Parsing Optimization

The worst was when I headed close to 40 good people in Oracle Premium Services in the late 90's, and I wanted them all to become OCP's: The Performance part of it was completely, utterly wrong and rubbish (based on 8i, as I recall it).

There was nothing relating to time in it - and this after some years with Anjo's YAPP paper and others beginning to wake up to time-based performance handling...

My guys (and girls) had to unlearn what they had learnt (we were damn proud in Premium of being among the first to use the time-based method) before going to the exam, then put their head back on again when they returned. It was not something they kept quiet about, for sure!

I think it began to get a little (very little) better in 9i, and in 10g it looks acceptable - the test, that is.

Mogens

Anjo Kolk wrote:

> Well
> I find every OCP question a trick question. Some times you have to
> answer the question wrong to get it right :-)
>
>
> On 9/1/06, *Polarski, Bernard* <Bernard.Polarski_at_atosorigin.com
> <mailto:Bernard.Polarski_at_atosorigin.com>> wrote:
>
> A bit OT:
>
> That would be a marvelous trap/question in the OCP exams, much
> clever than the grammatical we usually get.
>
> B. Polarski
>
>
> ------------------------------------------------------------------------
> *From:* Bobak, Mark [mailto:Mark.Bobak_at_il.proquest.com
> <mailto:Mark.Bobak_at_il.proquest.com>]
> *Sent: * Thursday, 31 August, 2006 6:32 PM
> *To:* Brandon.Allen_at_OneNeck.com
> <mailto:Brandon.Allen_at_OneNeck.com>; Polarski, Bernard;
> anjo.kolk_at_oraperf.com <mailto:anjo.kolk_at_oraperf.com>;
> ax.mount_at_gmail.com <mailto:ax.mount_at_gmail.com>;
> oracle-l_at_freelists.org <mailto:oracle-l_at_freelists.org>
> *Subject:* RE: 9i Parsing Optimization
>
>
> Right, session_cached_cursors can have an impact in reducing the
> amount of work a parse does, but, any time a piece of client code
> makes a parse call, Oracle MUST parse. It may be a hard parse, a
> soft parse, or a session cursor cache hit, but it is a parse, none
> the less. The *ONLY* way to reduce parsing is to change the
> program to parse less.
>
> As others speculated previously, perhaps there's an ODBC or JDBC
> driver, or something, that has an optimization that reduces the
> number of parse calls, that was picked up on the upgrade from 8i
> to 9i.....
>
> -Mark
>
> *--*
> *Mark J. Bobak*
> *Senior Oracle Architect*
> *ProQuest Information & Learning*
>
>
>
>
> --
> Anjo Kolk
> Owner and Founder OraPerf Projects
> tel: +31-577-712000
> mob: +31-6-55340888

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