Michael Glaesemann <grzm@seespotcode.net> writes: > I'd like to get confirmation that I'm correctly understanding the > times given in EXPLAIN ANALYZE. > ... > Is this correct? Looks about right to me. Note that some of what you are calling "executor overhead" might also be classed as "gettimeofday overhead". The measured difference between two successive gettimeofday readings presumably includes the actual userland runtime plus the equivalent of one gettimeofday call; but we actually did two calls. IOW the actual time to get in and out of a node is going to be a shade more than is reported. regards, tom lane ---------------------------(end of broadcast)--------------------------- TIP 6: explain analyze is your friend

Michael Glaesemann <grzm@seespotcode.net> writes: > I took startup time to be the time to return the first row *of the > first loop*. But it's actually the average startup time to return the > first row *in each loop*, right? Correct, just as the total time and tuples returned are averages over all the loops. regards, tom lane ---------------------------(end of broadcast)--------------------------- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to majordomo@postgresql.org so that your message can get through to the mailing list cleanly