At 15:29 +0200 4/19/02, Toomas Vendelin wrote:
>But RLIKE won't use indexes (even if they exist), so why RLIKE '^A' is
>better than LIKE 'A%', if LIKE which starts at the beginning of the
>string **will** use indexes?
I was comparing RLIKE '^A.*' to RLIKE '^A', not RLIKE to LIKE.
>PD> Regular expression patterns (unlike SQL patterns) don't need to match the
>PD> entire string, so RLIKE '^A' is sufficient and more efficient because
>PD> it doesn't spend time trying to match anything other than the first
>PD> As to the original question (below), that query looks like it should
>PD> work. Perhaps the problem lies elsewhere in the code that executes
>PD> the query. (It may be the query gets modified somehow. Without seeing
>PD> the context, it's hard to say.)