Re: Trigram performance penalty on varchar?
От | Alban Hertroys |
---|---|
Тема | Re: Trigram performance penalty on varchar? |
Дата | |
Msg-id | 9E5E1046-8144-4522-BC21-2511CEFC15C4@solfertje.student.utwente.nl обсуждение исходный текст |
Ответ на | Trigram performance penalty on varchar? (Mario Lopez <mario@lar3d.com>) |
Список | pgsql-general |
On Feb 21, 2008, at 10:56 AM, Mario Lopez wrote: > Hi, > > I am indexing a 100 million record table composed of varchar(255) > as the field to be indexed. I have always seen that examples of > pg_trgm are based on text type fields. Is this by any special reason?. A varchar is internally represented as text, with a size constraint of 255 characters in your case (I'm assuming your data requires that constraint?). > My computer is creating the index since 5 hours ago so I guess > there must be something wrong... I guess your server is running low on memory and the index being created doesn't fit in memory. You may want to have a look at http:// www.postgresql.org/docs/8.3/interactive/populate.html#POPULATE-WORK-MEM Another reason may be an exclusive lock on a row that you're trying to index, but that would mean that some transaction on some client somewhere is keeping that lock for a very long time (should not happen). You can check the pg_locks and pg_stat_activity tables for that. Regards, Alban Hertroys -- If you can't see the forest for the trees, cut the trees and you'll see there is no forest. !DSPAM:737,47c159af233092392031086!
В списке pgsql-general по дате отправления: