On Thu, Feb 11, 2010 at 7:48 AM, Bart Samwel <bart@samwel.tk> wrote:
> On Thu, Feb 11, 2010 at 13:41, Robert Haas <robertmhaas@gmail.com> wrote:
>>
>> On Thu, Feb 11, 2010 at 7:39 AM, Bart Samwel <bart@samwel.tk> wrote:
>> > Anyhow, I have no clue how much time the planner takes. Can anybody
>> > provide
>> > any statistics in that regard?
>>
>> It depends a great deal on the query, which is one of the things that
>> makes implementing this rather challenging.
>
> But I guess you can probably expect it to be on the same order for the same
> query in generic form and with filled-in parameters?
I think so.... but I wouldn't bet the farm on it without testing.
> Because that's the
> underlying assumption of the "ratio" criterion -- that re-planning with
> filled-in parameters takes about as much time as the initial planning run
> took.
We only want to replan when replanning is relatively cheap compared to
execution, so the other assumption is that the planning-to-execution
ratio is more or less constant. Whether that's sufficiently true to
make the proposed system useful and reliable is not clear to me.
...Robert