Re: Index Tuning Features - Mailing list pgsql-hackers

From Andrew Sullivan
Subject Re: Index Tuning Features
Date
Msg-id 20061011200021.GE25181@phlogiston.dyndns.org
Whole thread Raw
In response to Re: Index Tuning Features  (Gregory Stark <stark@enterprisedb.com>)
Responses Re: Index Tuning Features  (Theo Schlossnagle <jesus@omniti.com>)
List pgsql-hackers
On Wed, Oct 11, 2006 at 12:40:42PM -0400, Gregory Stark wrote:
> poorly written query. In fact Oracle is going in the opposite direction of
> even relying on hints internally. Its plan stability feature depends on
> generating and storing hints internally associated with every query.

But IBM, whose DB2 planner and optimiser is generally regarded as way
better than Oracle's (at least by anyone I know who's used both),
doesn't like hints.  The IBM people all say the same thing Tom has
said before: that the work to design the thing correctly is better
spent making the planner and optimiser parts smarter and cheaper,
because out of that work you also manage not to have the DBA
accidentally mess things up by simple-minded rule-based hints.  (Note
that I'm not trying to wade into the actual argument; I'm just
pointing out that even the biggest industry people don't agree on
this point.)

A

-- 
Andrew Sullivan  | ajs@crankycanuck.ca
A certain description of men are for getting out of debt, yet are
against all taxes for raising money to pay it off.    --Alexander Hamilton


pgsql-hackers by date:

Previous
From: "korryd@enterprisedb.com"
Date:
Subject: Re: hstore isexists
Next
From: "Jim C. Nasby"
Date:
Subject: Re: Clarification needed