Re: Big disappointment with Postgres
Date: Fri, 4 Feb 2011 20:48:30 +0000 (UTC)
Message-ID: <pan.2011.02.04.20.48.30_at_email.here.invalid>
On Fri, 04 Feb 2011 21:08:11 +0100, Thomas Kellerer wrote:
> Tom Kyte strongly detests hints as well:
>
> From:
> http://asktom.oracle.com/pls/apex/f?
p=100:11:0::::P11_QUESTION_ID:7038986332061
> Q: When should hints be used:
> A: Never. They are the path of last resort.
But they are here - as the path of the last resort.
>
>
> From:
> http://asktom.oracle.com/pls/apex/f?
p=100:11:0::::P11_QUESTION_ID:8912905298920
> "But -- I also have a rule -- don't use HINTS."
Tom is a technologist, not a DBA engaged in porting projects. Hints are bad, I grant you that, but the only thing worse than hints themselves is the absence of the hints when you need them. That is why every respectable RDBMS system in the world has them. Postgres is included here. It does have methods of controlling the optimizer, yet the "community" hypocritically insists on the idea that "hints are bad", instead of making them elegant and useful. In addition to that, there is nothing here that contradicts the argument of the correlated data. Argument of authority is a known logical fallacy and I am not sure that Tom Kyte would like seeing his name being used for that.
> I will leave this futile discussion now
BTW, this shows the willingness to really discuss the issues. There never was one. I was expected to bow down my head and accept the verdict of the gurus. Sorry, not me.
-- http://mgogala.byethost5.comReceived on Fri Feb 04 2011 - 14:48:30 CST