From: Jaime Casanova <systemguards@yahoo.com>
To: pgsql-performance@postgresql.org
Subject: Re: Sequential Scan with LIMIT
Date: Wed, 27 Oct 2004 12:16:21 -0500 (CDT)
Message-ID: <20041027171621.60824.qmail@web50006.mail.yahoo.com> (raw)
In-Reply-To: <417ECDEF.8030108@johnmeinel.com>
--- John Meinel <john@johnmeinel.com> escribió:
> Jaime Casanova wrote:
> [...]
> >>
> >>I'm not sure. They all return the same
> information.
> >
> >
> > of course, both queries will return the same but
> > that's just because you forced it.
> >
> > LIMIT and DISTINCT are different things so they
> behave
> > and are plenned different.
> >
> >
> >
> >>What's also weird is stuff like:
> >>SELECT DISTINCT(NULL) FROM mytable WHERE col =
> >>'myval' LIMIT 1;
> >
> >
> > why do you want to do such a thing?
> >
> > regards,
> > Jaime Casanova
> >
>
> I was trying to see if selecting a constant would
> change things.
> I could have done SELECT DISTINCT(1) or just SELECT
> 1 FROM ...
> The idea of the query is that if 'myval' exists in
> the table, return
> something different than if 'myval' does not exist.
> If you are writing a
> function, you can use:
>
> SELECT something...
> IF FOUND THEN
> do a
> ELSE
> do b
> END IF;
>
> The whole point of this exercise was just to find
> what the cheapest
> query is when you want to test for the existence of
> a value in a column.
> The only thing I've found for my column is:
>
> SET enable_seq_scan TO off;
> SELECT col FROM mytable WHERE col = 'myval' LIMIT 1;
> SET enable_seq_scan TO on;
>
> My column is not distributed well (larger numbers
> occur later in the
> dataset, but may occur many times.) In total there
> are something like
> 500,000 rows, the number 555647 occurs 100,000
> times, but not until row
> 300,000 or so.
>
> The analyzer looks at the data and says "1/5th of
> the time it is 555647,
> so I can just do a sequential scan as the odds are I
> don't have to look
> for very long, then I don't have to load the index".
> It turns out this
> is very bad, where with an index you just have to do
> 2 page loads,
> instead of reading 300,000 rows.
>
> Obviously this isn't a general-case solution. But if
> you have a
> situation similar to mine, it might be useful.
>
> (That's one thing with DB tuning. It seems to be
> very situation
> dependent, and it's hard to plan without a real
> dataset.)
>
> John
> =:->
>
In http://www.postgresql.org/docs/faqs/FAQ.html under
"4.8) My queries are slow or don't make use of the
indexes. Why?" says:
"However, LIMIT combined with ORDER BY often will use
an index because only a small portion of the table is
returned. In fact, though MAX() and MIN() don't use
indexes, it is possible to retrieve such values using
an index with ORDER BY and LIMIT:
SELECT col
FROM tab
ORDER BY col [ DESC ]
LIMIT 1;"
So, maybe you can try your query as
SELECT col FROM mytable
WHERE col = 'myval'
ORDER BY col
LIMIT 1;
regards,
Jaime Casanova
_________________________________________________________
Do You Yahoo!?
Información de Estados Unidos y América Latina, en Yahoo! Noticias.
Visítanos en http://noticias.espanol.yahoo.com
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-performance@postgresql.org
Cc: systemguards@yahoo.com
Subject: Re: Sequential Scan with LIMIT
In-Reply-To: <20041027171621.60824.qmail@web50006.mail.yahoo.com>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox