Re: enable_joinremoval - Mailing list pgsql-hackers

From Alex Hunsaker
Subject Re: enable_joinremoval
Date
Msg-id 34d269d41003291254p1e42b3a8n8cab1f1df04252db@mail.gmail.com
Whole thread Raw
In response to Re: enable_joinremoval  (Simon Riggs <simon@2ndQuadrant.com>)
Responses Re: enable_joinremoval
Re: enable_joinremoval
List pgsql-hackers
On Mon, Mar 29, 2010 at 13:41, Simon Riggs <simon@2ndquadrant.com> wrote:
>> Uhh... they wont be in the explain output...  Seems simple enough.
>
> That is exactly what I replied, though nobody felt that was a great
> answer.

Maybe I missed it, but why exactly do they care?  About the only
reason I can think *i* would care is:  If I was running the same SQL
on multiple database products (like mysql) and thought "Sweet now I
can use this new feature to cleanup my sql so it runs better on
product X or version of postgres Y".  Is there some other reason other
than it would be _cool_ to know?  Or is it FUD that it might be buggy
and so they wish to be able to turn it off?

It comes to mind you can probably do this with an plannerhook
(whatever happened to those hooks to auto create/recommend indexes?)


pgsql-hackers by date:

Previous
From: Dimitri Fontaine
Date:
Subject: Re: Proposal: Add JSON support
Next
From: Tom Lane
Date:
Subject: Re: enable_joinremoval