Re: PseudoPartitioning and agregates

Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:

> Greg Stark <gsstark(at)mit(dot)edu> writes:
>
> > How hard would it be to have Postgres actually remove the gettimeofday
> > overhead from the EXPLAIN ANALYZE output?
>
> Personally, I dislike measurement tools that lie to you under the flag
> of producing more-easily-interpreted results.

This is pretty standard practice for profilers in other contexts.

> As an example of why this would be a bad idea, the total time would no
> longer be closely related to the actual elapsed time (as measured by
> psql's \timing for instance) so you would be entirely unable to tell
> whether there was some significant factor not being measured.

Well that would be easily remedied by printing the total overhead subtracted
from all the nodes after the plan.

--
greg