Re: What happens between end of explain analyze and end of query execution ?

Le 05/04/2013 16:18, Nikolas Everett a écrit :
> On Fri, Apr 5, 2013 at 9:55 AM, Franck Routier
> <franck(dot)routier(at)axege(dot)com <mailto:franck(dot)routier(at)axege(dot)com>> wrote:
>
> Le 04/04/2013 21:08, Tom Lane a écrit :
> Maybe the statistics tables for sandia and saneds are in a bad
> shape ? (don't know how to check this).
>
> Regards,
>
> Franck
>
>
>
> Could this be caused by system table bloat?
>
> Also, can you check how long it takes to plan:
> 1. A query without a table at all (SELECT NOW())
> 2. A query with an unrelated table
>
> Again, what version of PostgreSQL is this?

PG 8.4.8
select now() is quick (15ms from pgadmin)
But I can reproduce the problem on other tables : explain on a query
with 4 join takes 4.5 seconds on an other set of tables...

System bloat... maybe. Not sure how to check. But as Tom asked,
default_statistics_target is 5000. Maybe the problem is here ? What
should I look after in pg_statistic to tell is there is a prolem here ?

Regards,
Franck