Re: 8.3devel slower than 8.2 under read-only load - Mailing list pgsql-hackers

From Guillaume Smet
Subject Re: 8.3devel slower than 8.2 under read-only load
Date
Msg-id 1d4e0c10711261104h63d6ced7i8a8c731ef9f93c18@mail.gmail.com
Whole thread Raw
In response to Re: 8.3devel slower than 8.2 under read-only load  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: 8.3devel slower than 8.2 under read-only load  ("Kevin Grittner" <Kevin.Grittner@wicourts.gov>)
List pgsql-hackers
On Nov 26, 2007 6:48 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> In the test case Guillame provided, every single WHERE clause happens
> to be of the form
>         varchar_column = 'unknown-type literal'
> and there are no other operators used in the SELECT lists; but I can
> hardly believe that this is representative of any significant number
> of real-world applications.  Even pgbench uses more than one operator.

Sure. The application uses a lot of other operators (timestamp
comparison, cube operators, LIKE, boolean comparison...). But as the
primary keys are all varchar in this application, the operator varchar
= unknown is by far the most used.

If we can't find a solution for upstream 8.3, I'll be happy to apply a
local patch for this customer (if I'm sure it doesn't do any harm in
their case).

Usually, people use integer flavoured primary keys so they probably
won't be hit by this problem as strong as we are.

--
Guillaume


pgsql-hackers by date:

Previous
From:
Date:
Subject: Re: Replacement Selection
Next
From: "Timothy J. Kordas"
Date:
Subject: Re: Replacement Selection