Re: *_collapse_limit, geqo_threshold

Alvaro Herrera <alvherre(at)commandprompt(dot)com> writes:
> Tom Lane escribi:
>> My own thought is that from_collapse_limit has more justification,
>> since it basically acts to stop a subquery from being flattened when
>> that would make the parent query too complex, and that seems like a
>> more understandable and justifiable behavior than treating JOIN
>> syntax specially.

> Isn't that what we use OFFSET 0 for? That one has also the nice
> property that you can actually specify which subquery you want to
> prevent from being flattened.

Well, if you want to modify your queries to prevent long planning times,
that'd be one way to do it. It doesn't seem like a generally useful
answer to me though. For example, typically the subquery would actually
be a view that might be used in various contexts. If you stick an
OFFSET in it then you disable flattening in all those contexts, likely
not the best answer.

> Personally I have never seen a case where the collapse_limits were
> useful tools.

I'm not convinced they're useful either.

regards, tom lane