Re: one more performance question I cannot test yet :(

From: Alvaro Herrera <alvherre(at)commandprompt(dot)com>
To: Sergei Shelukhin <realgeek(at)gmail(dot)com>
Cc: pgsql-general(at)postgresql(dot)org
Subject: Re: one more performance question I cannot test yet :(
Date: 2007-04-16 18:20:22
Message-ID: 20070416182022.GO5658@alvh.no-ip.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Sergei Shelukhin escribió:
> On Apr 15, 12:13 am, "Sergei Shelukhin" <realg(dot)(dot)(dot)(at)gmail(dot)com> wrote:
> > Suppose there's a table with two text fields and a lot of int fields.
> > THe table will be huge, updated constantly, and will be searched by
> > all of the int fields, whereas text fields will be for display only.
> >
> > Does it make sense to split it vertically into two tables, one with
> > ints and one with texts, and join them in queries to make it faster?
>
> Also, with a limited test data I have, I tried running queries against
> two table scheme...
> ifyou join two tables by id = id ( both ids are primary keys) it uses
> hash join, however as soon as you add a condition on one of the
> columns in ints' table, it degrades to nested loop. Wtf? Is there any
> way to force it to used clustered indexes for gods sake?!

No, you cannot force it. What you usually do is make sure it
understands that the indexed access is faster.

Have you run ANALYZE on the tables?

--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Tom Lane 2007-04-16 18:28:57 Re: PITR - "Rewind to snapshot" scheme
Previous Message Mikko Partio 2007-04-16 15:41:20