On Sun, Jul 15, 2001 at 10:20:33AM +0800, Lincoln Yeoh wrote:
> However I'm wondering what happens when a single query is fast, but you are
> doing thousands of it in actual use? Would the time resolution be
> sufficient in that case to point out the critical parts? This is just a
> very minor consideration tho.
For very short queries, the overhead of the connection and other things
happening on the machine will have far more effect than the plan chosen. I'm
aiming at the bigger stuff here.
> Another thing - if it actually runs the query, things might change in the
> database, so you might have to use a different name for the feature instead
> of EXPLAIN. If not people could screw things up without knowing it. Doing a
> rollback might help, but normally when users are already in a transaction
> they might not expect things like EXPLAIN to rollback the transaction for
> them.
Good point. I could limit it to SELECT statements ofcourse. All the other
things are select statements with some postprocessing.
--
Martijn van Oosterhout <kleptog@svana.org>
http://svana.org/kleptog/
> It would be nice if someone came up with a certification system that
> actually separated those who can barely regurgitate what they crammed over
> the last few weeks from those who command secret ninja networking powers.