On Mon, Mar 1, 2010 at 3:12 PM, Jean-Christophe Deschamps <j...@q-e-d.org> wrote: >>NULL = 12345 is NULL, NOT NULL is NULL, so subset N is not part of NOT >>(col = 12345). > > You're right of course! (and I was even saying about nulls treated apart) > > But, in your view, that the set can be non-contiguous for > negative/negated conditions would it explain that current code can't > make use of index?
I haven't been able to think of how it would preclude using the index, but I suspect it's more a matter of needing a similar-but-different codepath to optimize for the NOT case, rather than a simple "invert this" codepath relying on the existing case. Which is really just another way of stating Jay's point, perhaps. -scott _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users