Re: more on delete from join

From: Tegiri Nenashi <tegirinenashi_at_gmail.com>
Date: Thu, 27 Aug 2009 07:44:28 -0700 (PDT)
Message-ID: <b8c2256c-71d3-41d6-bf29-9a518245f94f_at_12g2000pri.googlegroups.com>


On Aug 27, 2:24 am, Nilone <rea..._at_gmail.com> wrote:
> My biggest frustration with views in existing products is not that
> some views aren't updateable, but that it is difficult to see whether
> a view is updateable or not, based on the syntax of query.

Which is not really product's fault. It is not easy to see if a query is reversible (that is if the input relation can be expressed in terms of the output) in Relational Algebra, let alone SQL. Arguably, Relational Lattice is the most algebraic-friendly system of the three, so that one can copy-and paste the query into some automatic theorem prover system (such as Prover9). Yet, the problem is so computationally complex that only relatively simple problems can be handled this way. Received on Thu Aug 27 2009 - 16:44:28 CEST

Original text of this message