Re: Query timing increased from 3s to 55s when used as function instead of select - Mailing list pgsql-performance

From Craig Ringer
Subject Re: Query timing increased from 3s to 55s when used as function instead of select
Date
Msg-id 4BFE95E1.4070701@postnewspapers.com.au
Whole thread Raw
In response to Re: Query timing increased from 3s to 55s when used as function instead of select  (Craig Ringer <craig@postnewspapers.com.au>)
List pgsql-performance
On 27/05/2010 11:33 PM, Craig Ringer wrote:
> On 21/05/2010 9:54 PM, Tyler Hildebrandt wrote:
>> We're using a function that when run as a select statement outside of the
>> function takes roughly 1.5s to complete whereas running an identical
>> query within a function is taking around 55s to complete.
>>
>> We are lost as to why placing this query within a function as opposed to
>> substituting the variables in a select statement is so drastically
>> different.
>
> This is a frequently asked question. It's the same issue as with
> prepared queries, where the planner has to pick a more general plan when
> it doesn't know the value of a parameter. The short answer is "work
> around it by using EXECUTE ... USING to invoke your query dynamically".
>
> ( Oddly, this FAQ doesn't seem to be on the FAQ list at
> http://wiki.postgresql.org/wiki/FAQ )

Added as:

http://wiki.postgresql.org/wiki/FAQ#Why_is_my_query_much_slower_when_run_as_a_prepared_query.3F

and the subsequent entry too.

Comments, edits, clarification appreciated. I know it's not as well
written as it could be, could use archive links, etc; it's just pass 1.


--
Craig Ringer

pgsql-performance by date:

Previous
From: David Jarvis
Date:
Subject: Re: Random Page Cost and Planner
Next
From: alvherre
Date:
Subject: Re: Autovacuum in postgres.