Re: Review remove {join,from}_collapse_limit, add enable_join_ordering - Mailing list pgsql-hackers

From Andres Freund
Subject Re: Review remove {join,from}_collapse_limit, add enable_join_ordering
Date
Msg-id 200907161518.02401.andres@anarazel.de
Whole thread Raw
In response to Re: Review remove {join,from}_collapse_limit, add enable_join_ordering  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: Review remove {join,from}_collapse_limit, add enable_join_ordering
Re: Review remove {join,from}_collapse_limit, add enable_join_ordering
List pgsql-hackers
On Thursday 16 July 2009 15:13:02 Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > The queries on the second reporting schema unfortunately are different.
> > Its the one were I copied the crazy example I attached in the original
> > thread. With geqo=off a good part of the queries used daily use too much
> > memory to plan sensibly and geqo=on outright fails with:
> > "Error: Failed to make a valid plan"
> > on some.
> We're not going to be able to fix this unless you show us examples.
In the other thread I attached a similar to the real schema + example query. 
Not enough? And why?

> > Noticeable even some plans which were plannable in reasonable time before
> > now are problematic with enable_join_ordering=false!
> And this even more so --- it doesn't make any sense at all.
Why? With a high from_collapse_limit more subqueries get inlined - before 
inlining they get planned separately.

> > So, while I think the changes are principally a good idea, as
> > {from,join}_collapse_limit are a bit confusing options, I personally! do
> > not think geqo is ready for it today, especially as the benefit is
> > relatively small.
> In general I think that any such bugs are there anyway and need to be
> fixed anyway.
Understandable.

Andres


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Review remove {join,from}_collapse_limit, add enable_join_ordering
Next
From: Tom Lane
Date:
Subject: Re: boolean in C