> The not-in subselect could be expensive, depending on table size. If > it turns out to be expensive, you might consider reference counting.
Thank you for the hint. It's done on an integer column with an index, so hopefully won't be too bad. Further tests will show though of course. Dennis _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
