Re: The Theoretical Foundations of the Relational Model
Date: Thu, 22 Aug 2002 21:28:47 GMT
Message-ID: <jIc99.5595$HY3.1518217_at_newsc.telia.net>
> Of course, order
> is allowed in result sets. Why should a result set be able to have
> some property (order, nonuniqueness of members) that base tables and
> relations do not?
I haven't read all in this thread, so I might miss out on something...
A resultset isn't ordered in RM. A resultset is a table and should have all the properties of a table. Not even the language SQL allows ordering in a result set (you are only allowed ORDER BY in a cursor definition - where we no longer have a resultset/table). Implementations (SQL Server, for instance) violates this, but ANSI SQL is clear on this subject.
The language SQL allows duplicates, and several feel that this (among other things) was a mistake. RM does not.
What I could agree on is how we derive the PK. There has probably been written plenty of the subject, things I haven't read. So I just leave it with this vague comment.
-- Tibor KarasziReceived on Thu Aug 22 2002 - 23:28:47 CEST