Re: query optimization scenarios 17,701 times faster!!!

From: Kurt Roeckx <Q(at)ping(dot)be>
To: Robert Dyas <rdyas(at)adelphia(dot)net>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: query optimization scenarios 17,701 times faster!!!
Date: 2003-04-24 18:07:44
Message-ID: 20030424180744.GA27347@ping.be
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Wed, Apr 23, 2003 at 02:04:02PM -0400, Robert Dyas wrote:
>
> The following is a list of query pairs (one fast, one slow) that must
> produce identical results by definition (and do), but have very different
> execution times.

I think what you see is what is described in the documentation
too. If you use outer joins, it will do them in the order of the
query. I think the reason for that was that the SQL standard
required it.

So if you put your query in an other order, you will get a
different result.

I believe that they removed that restriction in the last/cvs
version of PosgresQL.

> Especially the last example.

Where it's joining alot of tables it doesn't use in the first
place. You even removed the conditions on how to join the
tables.

Kurt

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Alice Lottini 2003-04-24 18:19:00 table creation using backend functions
Previous Message ohp 2003-04-24 18:00:07 Re: pg_dump dumping... core