Re: Can relational alegbra perform bulk operations?

From: Banana <Banana_at_Republic.com>
Date: Tue, 29 Sep 2009 14:03:02 -0700
Message-ID: <4AC27606.1080602_at_Republic.com>



Clifford Heath wrote:
> The theory allows the query to be reformulated without changing its
> meaning, but it cannot help estimate the cost of executing one or the other. A
> simpler expression may in fact cost more to evaluate.

Fascinating. It sounds like C.J. Date may have been too optimistic in asserting that relational theory is sufficient in optimizing the evaluation of a given expression. (well, he didn't say it can do so exclusively, but my puzzlement definitely dealt with the 'if relational theory is supposed to help with this, then why has SQL strayed so far from the relational theory as understood in mathematics.')

On an aside, though, that didn't quite answer my intended questions of whether theory provides a means of evaluating in bulk without requiring one to evaluate tuples individually, as I attempted to explain/ask in my other posts. But I suspect even if the answer was a yes, the theory has little to say about how one can optimize such queries.

> The change was triggered in part at least by the work of Ken J McDonell
> on the MUSBUS benchmarking tools, who was able to point out the glaring
> differences in actual behaviour of the RDBMS he benchmarked.

Thank you. I may look his works up and learn a bit about that. It does provide a hint (?) toward why the actual implementations is quite divorced from the relational theory. Received on Tue Sep 29 2009 - 16:03:02 CDT

Original text of this message