Here is the latest version of the EXPLAIN ANALYZE patch. I hope I correct
most of the issues that people had with the previous version. The impact
should be negligable when nothing is being measured. The precision is pretty
darn good considering its use of gettimeofday().
The effect is to create a new command EXPLAIN ANALYZE which actually
executes the query and then returns statistics on how the query ran. It
counts the number of rows, the time till the first tuple is returned, time
forall the tuples and the number of times the tuple was restarted.
If the node is part of a subplan, the statistics will be the total of all
executions of that subplan. I do beleive this is the right way of doing it.
Group nodes look slightly strange but I beleive it's fine that way.
Anyway, test away. If there is enough interest I can even write some docs.
It adds two files.
http://svana.org/kleptog/pgsql/timing-explain.patch
--
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.