Re: Why is outer Join way quicker?

From: Stephan Szabo <sszabo(at)megazone23(dot)bigpanda(dot)com>
To: David Link <dlink(at)soundscan(dot)com>
Cc: "pgsql-general(at)postgresql(dot)org" <pgsql-general(at)postgresql(dot)org>
Subject: Re: Why is outer Join way quicker?
Date: 2002-04-24 21:53:24
Message-ID: 20020424144756.U42332-100000@megazone23.bigpanda.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Wed, 24 Apr 2002, David Link wrote:

> Stephan Szabo wrote:
> >
> > See:
> > http://www.postgresql.org/idocs/index.php?explicit-joins.html
> >
> > You can probably use an inner join as well, you just need to use
> > the explicit join syntax.
>
> OK. I read that and it helps explain this behavior (new to 7.1)
> somewhat. Thank you.
>
> However, it also says that the planner spends some time figuring out
> which method of joining is best (a join (b join c)) vs. ((a join b) join
> c), etc. It says the user should not have to worry about explicitly
> mapping the join order when few tables are involved. In this case I am
> using three tables, though fairly large. See explain plans.
>
> You'll notice the explain plans show a difference of magnatude in cost.
> If the planner is checking those kind of things, it should pick up on
> this and choose the better join order.

You are explicitly mapping the join order in both queries. I'm not 100%
sure the page mentions, but I think it does... If you do
a,b OUTER JOIN c, postgres takes that as a request to explicitly join
b and c before any other joins with those tables, thus any plan that
joins a to b or c before the outer join is not considered. The thing for
explicitly mapping join order is because with a large number of tables
the decision making process can be complicated and take a fair amount of
time, so you can prune the number of possible join orders with explicitly
using the sql92 join syntax.

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Tom Lane 2002-04-24 22:08:43 Re: temp space on different location
Previous Message PG Explorer 2002-04-24 21:52:30 Re: odd psql behaviour on OSX