Re: Operator performance question

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Alban Hertroys <alban(at)magproductions(dot)nl>
Cc: Postgres General <pgsql-general(at)postgresql(dot)org>
Subject: Re: Operator performance question
Date: 2007-01-09 16:31:20
Message-ID: 8233.1168360280@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Alban Hertroys <alban(at)magproductions(dot)nl> writes:
> My conclusion is that this query time is mostly limited to the somewhat
> complex COUNT expressions. Is there any way to do this more efficiently?

Offhand I would bet on the bitstring-AND operations being the
bottleneck; you could test this by comparing the speed of queries that
are doing different mixes of the same number of COUNT()s. If you're
happy with a fixed-width 32-bit field, consider using an integer field
and integer & operations, instead of bitstring. Bitstring is a
pass-by-reference type and so inherently a lot less efficient than an
integer.

Another suggestion is to replace

count(nullif(boolean_expr, false))

with

sum((boolean_expr)::int)

I think this would be a marginal speed win at best (basically replacing
a Const and a NullIf node with a Cast node), but it just seems to me
to be more natural ... it took me a bit to figure out what your query
was trying to accomplish.

regards, tom lane

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Matthew T. O'Connor 2007-01-09 16:31:59 Re: Autovacuum Improvements
Previous Message Brandon Aiken 2007-01-09 16:24:57 Re: Operator performance question