On a large table it will be much faster to use an index than to force a row scan as your initial query did.
L B wrote: > I have obtained a great performance improvement now > just adding an index in the 2 columns > > dtfrom and idcatalogue, > > removing the 2 single indexes on the column dtfrom and > idcatalogue. > > I could not believe, now the query > > SELECT MAX(DTFROM) AS MAXDT, IDCATALOGUE FROM > CRONOCATALOGUES GROUP BY IDCATALOGUE > > takes only 2 seconds in sqlite and 5 seconds in > access. > May be the field idcatalogue has not to be found in > the table anymore because it is in the index (this is > my explanation, but I'm not sure). > Could not imagine such an improvement! > > Sorry for my ignorance. > > > > > _______________________________________________ > 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
