On Thu, Oct 06, 2005 at 07:29:44PM -0400, Tom Lane wrote: > Well, it's surely legitimate material as a "gotcha". The example is > taken from > http://archives.postgresql.org/pgsql-general/2004-11/msg01375.php > and the "previous discussion" referred to is this thread: > http://archives.postgresql.org/pgsql-hackers/2004-10/msg00082.php > > As you can see, there wasn't a lot of agreement that we ought to change > it. Arguably, if we did change it we'd get ripped for the "gotcha" of > poor optimization when the user forgets to mark nonvolatile functions > properly. (Personally, though, I'm in favor of tightening it up.) Aren't there a good number of performance issues if you mis-mark a function? In any case, ISTM it'd be much better to perform poorly rather than give bad/wrong results. I don't really see much dissention there, so I'd vote for making the change for 8.2. -- Jim C. Nasby, Sr. Engineering Consultant jnasby@xxxxxxxxxxxxx Pervasive Software http://pervasive.com work: 512-231-6117 vcard: http://jim.nasby.net/pervasive.vcf cell: 512-569-9461 ---------------------------(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
