Here is his clarifier...
Maybe someone might want to get a case study out of him...
Chris
----- Original Message -----
From: "Eirik Oeverby" <ltning@anduin.net>
To: "Christopher Kings-Lynne" <chriskl@familyhealth.com.au>
Sent: Tuesday, July 08, 2003 2:43 PM
Subject: Re: IN list processing performance (yet again)
Hi,
It's ok, but please do be fair and mention that the datasets aren't 100%
equal - the tables are narrower (in the postgres version I've stripped
out columns that aren't needed), but the fields worked on in those
particular queries are the same. And the cardinality is comparable.
Another interesting thing to note is that the postgres database is
running on weaker hardware than the DB2 server - which I suppose is
giving credit to the less bloated nature of postgres.
The conclusion is, however, that this is a database that - when we set
it up 2.5 years ago - could only be run on DB2. All other databases we
tested (MS SQL, Oracle, Postgres, MySQL) were unacceptable either
performance-wise or feature-wise. Postgre has now (with 7.4-CVS ;)
matured enough to be a serious alternative - which is a very Good Thing
(TM).
Thanks again for your help!
(And yes, I am backing up every night. ;)
/Eirik
On Tue, 8 Jul 2003 09:42:58 +0800
"Christopher Kings-Lynne" <chriskl@familyhealth.com.au> wrote:
> Hi Eirik,
>
> Would you mind if I quoted you on the advocacy list with regards to
> your'faster than DB2 for IN processing' experience?
>
> Chris
>
> ----- Original Message -----
> From: "Eirik Oeverby" <ltning@anduin.net>
> To: "Christopher Kings-Lynne" <chriskl@familyhealth.com.au>
> Sent: Monday, July 07, 2003 8:35 PM
> Subject: Re: IN list processing performance (yet again)
>
>