Re: [PERFORM] Measuring execution time for sql called from PL/pgSQL

2003-12-16 Thread David Shadovitz
> I've tried to measure the duration of sql with printing out > "localtimestamp" but for some reason during the same pg/plsql call > it returns the same value: Aram, >From http://www.postgresql.org/docs/current/static/functions-datetime.html: There is also the function timeofday(), which for h

Re: [PERFORM] Measuring execution time for sql called from PL/pgSQL

2003-12-12 Thread Tomasz Myrta
Dnia 2003-12-12 02:17, Użytkownik Aram Kananov napisał: select localtimestamp into v; raise notice ''Timestamp: %'', v; Don't use localtimestamp, now() neither any transaction based time function. They all return the same value among whole transaction. The only time function, which can be

[PERFORM] Measuring execution time for sql called from PL/pgSQL

2003-12-11 Thread Aram Kananov
Hi, I've got very slow insert performance on some table which has trigger based on complex PL/pgSQL function. Apparently insert is slow due some slow sql inside that function, since CPU load is very high and disk usage is low during insert. I run Red Hat 9 Anthlon 2.6 1GB ram Fast IDE Disk Setti