Re: selecting a column according to a minimum

From: Mikito Harakiri <mikharakiri_nospaum_at_yahoo.com>
Date: 20 Oct 2004 09:24:33 -0700
Message-ID: <8a529bb.0410200824.93192ef@posting.google.com>

Serge Rielau <srielau_at_ca.ibm.com> wrote in message news:<2tlvqsF21a6qfU1_at_uni-berlin.de>...
> Correct. The DB2 II development team obviosuly has Oracle installed to
> write and test wrappers.
> Not needing to know I have no access to those machines.
> If I had O10g on my laptop no one would belive me that I don't look at
> the optimizer plans. Plans tell a lot about the optimizer and query
> rewrite. "Reverse engineering"
> If I ran a query on O10g and I run the same query on DB2 I would have a
> hell of a time controlling my tongue (assuming DB2 would be faster ;-)
> "publishing a benchmark without Oracle's consent".
> Lawyers are no fun at all :-(

This is all gray area, I agree.

As far as optimizer is concerned, execution plan tells you only one part of the story: physical optimizer. Query transformation/rewrite is arguably much more sophisticated part and is not reflected an plan table or trace file. BTW, does DB2 exposes the decision tree of transformations somehow? Received on Wed Oct 20 2004 - 11:24:33 CDT