Re: The use of cpu_index_tuple_cost by the query planner - Mailing list pgsql-novice

From Antonio Carlos Salzvedel Furtado Junior
Subject Re: The use of cpu_index_tuple_cost by the query planner
Date
Msg-id CAERqmVq5xO5-JR595FDkmPT3i2xP306NVKuW-zftUC7QRSJSaw@mail.gmail.com
Whole thread Raw
In response to Re: The use of cpu_index_tuple_cost by the query planner  (Jeff Davis <pgsql@j-davis.com>)
List pgsql-novice
Thanks,

I had to look in the soure code for the other parameters as well.  I was trying to create some queries because I was interested in the relation between estimated cost and actual time.

Regards,

Antonio Carlos Furtado

On Wed, Jun 27, 2012 at 8:22 PM, Jeff Davis <pgsql@j-davis.com> wrote:
On Wed, 2012-06-27 at 21:37 +0100, Simon Riggs wrote:
> On 27 June 2012 21:28, Antonio Carlos Salzvedel Furtado Junior
> <acsfj08@inf.ufpr.br> wrote:
>
> > I'm trying to understand the use of PostgreSQL tuning parameters by the
> > query planner's cost estimator. I'm trying to use simple queries to
> > understand how these parameters would affect the estimated cost. However, I
> > haven't still been able to see the cpu_index_tuple_cost.
>
> Your best hope is to read the source code. Things will become much clearer.

Right, the cost model is too complex to really understand just running
EXPLAIN.

The tuning parameters are translated to global variables
in ./src/backend/utils/misc/guc.c. From there, you can search for
references to the global variables (often with very similar names to the
tuning parameter) and find out how they are used in the cost model.

Regards,
       Jeff Davis


pgsql-novice by date:

Previous
From: Jeff Davis
Date:
Subject: Re: Function Scan costs
Next
From: Bob Holmstrom
Date:
Subject: Problem installing PostGIS because of pg_config