Re: Rewrite, normal execution vs. EXPLAIN ANALYZE

On 7/23/2010 10:46 PM, Andres Freund wrote:
> On Fri, Jul 23, 2010 at 03:30:03PM -0400, Robert Haas wrote:
>> On Fri, Jul 23, 2010 at 3:19 PM, Marko Tiikkaja
>> <marko(dot)tiikkaja(at)cs(dot)helsinki(dot)fi> wrote:
>>> This may be a bit hard to follow, but essentially what happens is that in
>>> EXPLAIN ANALYZE, the INSERT in the rule does not see the changes made by T2
>>> to baz while in the regular execution scenario it does.
>>
>> Well that's gotta be a bug, but in what I'm not sure.
> One could argue that its less of a semantic change changing explain's
> behaviour than the normal executors way of working...

One could also argue that people usually test their code with EXPLAIN
ANALYZE and could've made the opposite conclusion based on its output. ;-)

But I really have no idea what we should do about this. It seems to me
that EXPLAIN ANALYZE's behaviour is less surprising (that's actually
what I, before yesterday, always thought happens), but I'm not too sure
about that either.

Regards,
Marko Tiikkaja