Re: [PERFORM] text search: tablescan cost for a tsvector

2012-02-29 Thread Robert Haas
On Mon, Feb 6, 2012 at 6:05 AM, Marc Mamin m.ma...@intershop.de wrote: without analyze: http://explain.depesz.com/s/6At with analyze:    http://explain.depesz.com/s/r3B I think this is the same issue complained about here: http://archives.postgresql.org/message-id/4ed68eec.9030...@krogh.cc

Re: [PERFORM] text search: tablescan cost for a tsvector

2012-02-29 Thread Marc Mamin
Von: Robert Haas [mailto:robertmh...@gmail.com] Gesendet: Mi 2/29/2012 7:32 On Mon, Feb 6, 2012 at 6:05 AM, Marc Mamin m.ma...@intershop.de wrote: without analyze: http://explain.depesz.com/s/6At with analyze:http://explain.depesz.com/s/r3B ... The problem seems to be that the

[PERFORM] text search: tablescan cost for a tsvector

2012-02-06 Thread Marc Mamin
Hello, I have quite systematically better performance with the text search when I disable the statistics collection for the tsvector column. So I wonder if such statistics ever make sense. Here a testcase: The table contains 200'000 tsvector, whereas the lexeme 'fooblablabla' exists in all