ve3meo <[email protected]> wrote: > If the number of VM instructions is a good inverse indicator of efficiency, > then the second query with the index appears to be the most efficient. I'm > sure somebody will point out if it is not.
It is not. It may very well take fewer instructions to implement a straightforward linear scan than it is to join through the index. -- Igor Tandetnik _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
