Quick note to self and others: IN() and EXISTS() in all RDB's I know of are the uncle noone mentions; that is to say, they have different compare semantics than JOIN so the naive strategy is to evaluate them as dependent subqueries, not correlated ones, which would be consistent with the behavior I noticed. However, I do know of Maria and Postgres that do a decent job at optimizing EXISTS () (which I think is by all means the correct semantic for this intent). But there's by no means a golden standard across RDB's so that's why it would be very useful to have some documentation on it, as it's one of the migration pitfalls.
I'm still in debt with the TC, will work on it the next days. -- Sent from: http://sqlite.1065341.n5.nabble.com/ _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
