Re: experimental pg_qcache patch - Mailing list pgsql-hackers

From Jean-Paul ARGUDO
Subject Re: experimental pg_qcache patch
Date
Msg-id 20020416082500.GA2875@pastis
Whole thread Raw
In response to experimental pg_qcache patch  (Neil Conway <nconway@klamath.dyndns.org>)
Responses Re: experimental pg_qcache patch
List pgsql-hackers
> I've attached an updated version of Karel Zak's pg_qcache patch, which
> adds PREPARE/EXECUTE support to PostgreSQL (allowing prepared SQL
> statements). 

Woah :-))

Thanks Neil! You may be remind of a thread in february, where I talked
about a survey about migrating from Oracle 8.0 / NT4 to PostgreSQL 7.2 /
Red Hat 7.2 ...

Overall performances obtained are a ratio of 1,33 on standard queries
of the application, like on migrated CONNECT BY Oracle statements
(thanks again to OpenACS guys for this). This ratio is very good for us
and our customer. We felt some pride about such results.

But we faced a problem in migrating bulk plain batch in Oracle Pro*C to
ECPG: performances where 3 times slower, due to incapacity of PG to
prepare statments (some well informed guys here in PG list gave us
tips&hints to use SPI's prepared statment. Unfortunately, this would
result in a loss of functionalities from Pro*C to ECPG.. :-( so we had
to abandon this issue).

I talk here about CURSORs. 

I imagine that with your patch, we could prepare statments used in
cursors. We going to test this and benchmark the application. Not sure
it works, I think ECPG has first to take into consideration those new
functialities (Michael?).

Be sure to have feedback on this :-)

Thanks again for such initiative! I'm going to inform my co-worker (C++
senior) on your patch with the hope he can help you.

Cheers,

-- 
Jean-Paul ARGUDO                                IDEALX S.A.S
Consultant bases de données                     15-17, av. de Ségur
http://www.idealx.com                           F-75007 PARIS


pgsql-hackers by date:

Previous
From: "Christopher Kings-Lynne"
Date:
Subject: Re: Places to update when a new version is out
Next
From: Tatsuo Ishii
Date:
Subject: Re: multibyte support by default