Re: Left Join Not Using Index?
От | Dennis Gearon |
---|---|
Тема | Re: Left Join Not Using Index? |
Дата | |
Msg-id | 3EA6B14C.4030406@cvc.net обсуждение исходный текст |
Ответ на | Re: Left Join Not Using Index? (Hunter Hillegas <lists@lastonepicked.com>) |
Список | pgsql-general |
I think he is saying that efforts in data design, and query design are called for. Hunter Hillegas wrote: > So, what you're basically saying is that my best bet for improving the speed > of a query to get this result set is to tweak my settings and get faster > hardware? > > >>From: Tom Lane <tgl@sss.pgh.pa.us> >>Date: Wed, 23 Apr 2003 01:30:05 -0400 >>To: Hunter Hillegas <lists@lastonepicked.com> >>Cc: Dann Corbit <DCorbit@connx.com>, Stephan Szabo >><sszabo@megazone23.bigpanda.com>, PostgreSQL <pgsql-general@postgresql.org> >>Subject: Re: [GENERAL] Left Join Not Using Index? >> >>Hunter Hillegas <lists@lastonepicked.com> writes: >> >>>Can anyone point me in another direction to optimize this >> >>AFAICS you cannot improve that without changing the query structure >>and/or the database layout. Because the WHERE clause is a bunch of OR'd >>conditions, it's useless for restricting either individual table scan >>making up the join --- there is really no implementation short of >>forming the entire join described by the FROM ... JOIN ... ON ... part >>of the query and then testing each individual row against the WHERE >>clause. While that is the abstract semantic model implied by the SQL >>spec, it's not exactly how you want a query to really be done :-(. >> >>regards, tom lane > > > > ---------------------------(end of broadcast)--------------------------- > TIP 6: Have you searched our list archives? > > http://archives.postgresql.org >
В списке pgsql-general по дате отправления: