|From:||Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>|
|To:||Matthew Wakeling <matthew(at)flymine(dot)org>|
|Cc:||Robert Haas <robertmhaas(at)gmail(dot)com>, Merlin Moncure <mmoncure(at)gmail(dot)com>, Grzegorz Jaśkiewicz <gryzman(at)gmail(dot)com>, pgsql-performance(at)postgresql(dot)org|
|Subject:||Re: Really dumb planner decision|
|Views:||Raw Message | Whole Thread | Download mbox | Resend email|
Matthew Wakeling <matthew(at)flymine(dot)org> writes:
> On Thu, 16 Apr 2009, Robert Haas wrote:
>> I hasten to point out that I only suggested raising them to the moon
>> as a DEBUGGING strategy, not a production configuration.
> The problem is that we have created a view that by itself a very
> time-consuming query to answer, relying on it being incorporated into a
> query that will constrain it and cause it to be evaluated a lot quicker.
> This kind of scenario kind of guarantees a bad plan as soon as the number
> of tables reaches from_collapse_limit.
Well, if the payoff for you exceeds the extra planning time, then you
raise the setting. That's why it's a configurable knob. I was just
pointing out that there are downsides to raising it further than
regards, tom lane
|Next Message||Matthew Wakeling||2009-04-16 16:06:03||GiST index performance|
|Previous Message||Tom Lane||2009-04-16 16:01:12||Re: Shouldn't the planner have a higher cost for reverse index scans?|