I'll have to look again. I was sure I had mucked with estimatedCost and found that it wasn't doing the trick for what I meant to do - but somehow the context has swapped out of my short-term memory, so I'll have to recreate it before I can say why this was the case.
-scott On Wed, Sep 10, 2008 at 1:43 PM, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > > see sqlite3_index_info.estimatedCost > > http://www.sqlite.org/cvstrac/chngview?cn=5649 > >> AFAICT, when you have a join where one table has a good index, the >> virtual table cannot signal that it has an even better index. I could >> not follow the index-selection logic well enough to have any >> suggestions how to improve things. In the fts case, this breaks the >> query because the MATCH operator _only_ works when used to access the >> index. >> >> -scott > > -- > View this message in context: > http://www.nabble.com/rtree-performance-problems--tp19035092p19422728.html > Sent from the SQLite mailing list archive at Nabble.com. > > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users