> Date: Mon, 28 Jan 2008 12:08:00 -0800 > From: [EMAIL PROTECTED] > To: [EMAIL PROTECTED] > CC: pgsql-patches@postgresql.org > Subject: Re: [PATCHES] Auto-explain patch > > On Mon, Jan 28, 2008 at 07:55:53PM +0000, Dean Rasheed wrote: >> >>> Dean, >>> >>> Maybe I missed something obvious here, but how does this patch handle >>> the situation where people have turned on INTEGER_DATETIMES? >>> >>> Cheers, >>> David. >>> -- >>> David Fetter http://fetter.org/ >>> Phone: +1 415 235 3778 AIM: dfetter666 Yahoo!: dfetter >>> Skype: davidfetter XMPP: [EMAIL PROTECTED] >>> >>> Remember to vote! >>> Consider donating to Postgres: http://www.postgresql.org/about/donate >> >> Sorry, I don't understand. I am new to this code, but I can't see >> how the INTEGER_DATETIMES flag will affect this code. I am using the >> macros and functions from instrument.h and explain.c for timing, >> which appear to use gettimeofday() or QueryPerformanceCounter(), >> coverting the result to a double to report the total time spent >> running the query. > > It's the double part I don't quite get. Shouldn't that be an int64 in > the case of INTEGER_DATETIMES? > > Cheers, > David. >
All the times are 64-bit integers (or at least structures with 2 32-bit integers in them) until the end, when the elapsed time is converted to a double so that the query runtime can be printed out in ms ("Query runtime: %.3f ms"). This is the same as EXPLAIN ANALYSE, except in that case it is the total runtime ("Total runtime: %.3f ms\n") that gets reported, including startup/shutdown trigger times. Dean. _________________________________________________________________ Share what Santa brought you https://www.mycooluncool.com ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match