Re: Determining which index to create
От | Martijn van Oosterhout |
---|---|
Тема | Re: Determining which index to create |
Дата | |
Msg-id | 20011122122112.A2666@svana.org обсуждение исходный текст |
Ответ на | Re: Determining which index to create (Eric Cholet <cholet@logilune.com>) |
Ответы |
Re: Determining which index to create
|
Список | pgsql-general |
On Wed, Nov 21, 2001 at 04:09:52PM +0100, Eric Cholet wrote: > => explain select * from dico_frs where motid=4742 order by date desc limit > 10; > NOTICE: QUERY PLAN: > > Limit (cost=46172.25..46172.25 rows=10 width=16) > -> Sort (cost=46172.25..46172.25 rows=11382 width=16) > -> Index Scan using dico_frs_motid_date on dico_frs > (cost=0.00..45405.39 rows=11382 width=16) That's wrong. It doesn't seem to realise that a reverse scan on the index would give the right answer. Note that that's only true because you're selecting only a single motid. If there were multiple, a reverse scan would definitly not be appropriate. Not sure how to fix this though. Functional index? -- Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/ > Magnetism, electricity and motion are like a three-for-two special offer: > if you have two of them, the third one comes free.
В списке pgsql-general по дате отправления: