Re: question about efficiency - Mailing list pgsql-general

From
Subject Re: question about efficiency
Date
Msg-id 49166.192.168.1.32.1037039664.squirrel@mainbox.archonet.com
Whole thread Raw
In response to Re: question about efficiency  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: question about efficiency  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general
> Stephan Szabo <sszabo@megazone23.bigpanda.com> writes:
>> It's basically converted it into a three way join.
>
> More to the point, an *unconstrained* join between "claimsum2001
> as claimsum2001" and the intended joined table.  The reason it
> ran for so long was it was generating a huge number of rows ---
> as many as would be in the cross product.

Would it be possible to have some sort of configurable limit on this:

max_result_set_size = 10,000,000
max_result_set_action = ignore | debug | warn | error

The planner knows how many rows it's expecting to return, and a warning
message would show up unconstrained queries like this.

- Richard Huxton



pgsql-general by date:

Previous
From: "scott.marlowe"
Date:
Subject: Re: Upgrade to dual processor machine?
Next
From: Josh Berkus
Date:
Subject: Re: Hardware estimation