On Mon, May 12, 2014 at 5:41 PM, Jan Slodicka <[email protected]> wrote:
> I think that the answer is as before: The LIKE optimization cannot be > performed by the DB engine, but only by the collation author. > It can be implemented if the definition of a collation is extended to be able to provide this information as Simon suggested. For example, by defining a second callback that provides the lower and upper bounds that enclose the results of a LIKE prefix string. If this callback is not implemented or if it returns null strings then no optimization should be attempted. _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
