could you provide me a dump of your table (just id and tsvector columns),
so I could try on my computer. Also, plain query (simple and clean) which
demonstrated your problem would be preferred next time !
Oleg
On Tue, 16 Nov 2004, [iso-8859-15] Herv? Piedvache wrote:
Michael,
Le Mardi 16 Novembre 2004 16:32, Michael Fuhr a ?crit :
On Tue, Nov 16, 2004 at 03:55:58PM +0100, Herv? Piedvache wrote:
WHERE s.idx_site_name @@ to_tsquery('atari');
How much text does each site_name field contain? From the field
name I'd guess only a few words. Based on my own experience, if
the fields were documents containing thousands of words then I'd
expect tsearch2 to be faster than ILIKE by an order of magnitude
or more.
Yes site name ... is company names or web site name ... so not many word in
each record ... but I don't understand why more words are more efficient than
few words ?? sorry ...
Regards,
Regards,
Oleg
_____________________________________________________________
Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
Sternberg Astronomical Institute, Moscow University (Russia)
Internet: oleg@xxxxxxxxxx, http://www.sai.msu.su/~megera/
phone: +007(095)939-16-83, +007(095)939-23-83
