pg.ddx.io pgsql-sql@postgresql.org mailing list archive
help / color / mirror / Atom feedoptimizing select ... not in (select ...)
5+ messages / 4 participants
[nested] [flat]
* optimizing select ... not in (select ...)
@ 2001-08-13 13:17 Laurent Martelli <laurent@aopsys.com>
0 siblings, 3 replies; 5+ messages in thread
From: Laurent Martelli @ 2001-08-13 13:17 UTC (permalink / raw)
To: pgsql-sql
I have this query :
select distinct on (Pictures.PictureID) * from Pictures where Pictures.PictureID not in (select distinct PictureID from Keywords);
and I find it a bit slow. Does anybody have suggestions to run this
faster ? (I have indexes on PictureID on both Pictures and Keywords)
--
Laurent Martelli
laurent@aopsys.com http://www.bearteam.org/~laurent/
^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: optimizing select ... not in (select ...)
@ 2001-08-13 14:55 Tomas Berndtsson <tomas@nocrew.org>
parent: Laurent Martelli <laurent@aopsys.com>
2 siblings, 1 reply; 5+ messages in thread
From: Tomas Berndtsson @ 2001-08-13 14:55 UTC (permalink / raw)
To: Laurent Martelli <laurent@aopsys.com>; +Cc: pgsql-sql
Laurent Martelli <laurent@aopsys.com> writes:
> I have this query :
>
> select distinct on (Pictures.PictureID) * from Pictures where Pictures.PictureID not in (select distinct PictureID from Keywords);
>
> and I find it a bit slow. Does anybody have suggestions to run this
> faster ? (I have indexes on PictureID on both Pictures and Keywords)
Try this instead:
select distinct on (Pictures.PictureID) *
from Pictures
where not exists (select distinct PictureID from Keywords where Pictures.PictureID=Keywords.PictureID);
I've found that it's usually faster, probably because it can use
indices better.
Tomas
^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: optimizing select ... not in (select ...)
@ 2001-08-13 15:16 Josh Berkus <josh@agliodbs.com>
parent: Laurent Martelli <laurent@aopsys.com>
2 siblings, 0 replies; 5+ messages in thread
From: Josh Berkus @ 2001-08-13 15:16 UTC (permalink / raw)
To: Laurent Martelli <laurent@aopsys.com>; pgsql-sql
Laurent,
> select distinct on (Pictures.PictureID) * from Pictures where
> Pictures.PictureID not in (select distinct PictureID from Keywords);
>
> and I find it a bit slow. Does anybody have suggestions to run this
> faster ? (I have indexes on PictureID on both Pictures and Keywords)
NOT IN is always slow on all RDBMS that I know, unless the subselect has
a very small (<100) return set. This is because the engine has to
compare each row in the master query against every value returned by NOT
IN, one row at a time.
Try the WHERE NOT EXISTS construction instead:
SELECT * FROM Pictures
WHERE NOT EXISTS ( SELECT pictureID FROM keywords
WHERE keywords.pictureID = Pictures.pictureID );
This uses the DB engine's JOIN functionality and thus runs considerably
faster.
BTW, all those "DISTINCT" in the query example you gave, assuming that
PictureID is the unique index of Pictures, are completely superfluous
and will only slow the query down. Particularly the use of DISTINCT in
a subquery should only be used if the contents of the subquery will be
displayed as part of the result set.
-Josh Berkus
______AGLIO DATABASE SOLUTIONS___________________________
Josh Berkus
Complete information technology josh@agliodbs.com
and data management solutions (415) 565-7293
for law firms, small businesses fax 621-2533
and non-profit organizations. San Francisco
^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: optimizing select ... not in (select ...)
@ 2001-08-13 15:59 Bruce Momjian <pgman@candle.pha.pa.us>
parent: Laurent Martelli <laurent@aopsys.com>
2 siblings, 0 replies; 5+ messages in thread
From: Bruce Momjian @ 2001-08-13 15:59 UTC (permalink / raw)
To: Laurent Martelli <laurent@aopsys.com>; +Cc: pgsql-sql
Can it be redone using NOT EXISTS?
> I have this query :
>
> select distinct on (Pictures.PictureID) * from Pictures where Pictures.PictureID not in (select distinct PictureID from Keywords);
>
> and I find it a bit slow. Does anybody have suggestions to run this
> faster ? (I have indexes on PictureID on both Pictures and Keywords)
>
> --
> Laurent Martelli
> laurent@aopsys.com http://www.bearteam.org/~laurent/
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
>
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026
^ permalink raw reply [nested|flat] 5+ messages in thread
* Re: optimizing select ... not in (select ...)
@ 2001-08-13 16:35 Laurent Martelli <laurent@aopsys.com>
parent: Tomas Berndtsson <tomas@nocrew.org>
0 siblings, 0 replies; 5+ messages in thread
From: Laurent Martelli @ 2001-08-13 16:35 UTC (permalink / raw)
To: pgsql-sql
>>>>> "Tomas" == Tomas Berndtsson <tomas@nocrew.org> writes:
Tomas> Laurent Martelli <laurent@aopsys.com> writes:
>> I have this query :
>>
>> select distinct on (Pictures.PictureID) * from Pictures where
>> Pictures.PictureID not in (select distinct PictureID from
>> Keywords);
>>
>> and I find it a bit slow. Does anybody have suggestions to run
>> this faster ? (I have indexes on PictureID on both Pictures and
>> Keywords)
Tomas> Try this instead:
Tomas> select distinct on (Pictures.PictureID) * from Pictures where
Tomas> not exists (select distinct PictureID from Keywords where
Tomas> Pictures.PictureID=Keywords.PictureID);
Tomas> I've found that it's usually faster, probably because it can
Tomas> use indices better.
It is indeed much faster. Thanks to all of you who answered so fast.
--
Laurent Martelli
laurent@aopsys.com http://www.bearteam.org/~laurent/
^ permalink raw reply [nested|flat] 5+ messages in thread
end of thread, other threads:[~2001-08-13 16:35 UTC | newest]
Thread overview: 5+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2001-08-13 13:17 optimizing select ... not in (select ...) Laurent Martelli <laurent@aopsys.com>
2001-08-13 14:55 ` Tomas Berndtsson <tomas@nocrew.org>
2001-08-13 16:35 ` Laurent Martelli <laurent@aopsys.com>
2001-08-13 15:16 ` Josh Berkus <josh@agliodbs.com>
2001-08-13 15:59 ` Bruce Momjian <pgman@candle.pha.pa.us>
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