Re: 4 the FAQ: Are Commercial DBMS Truly Relational?

"Kenneth Downs" <firstinit.lastname_at_lastnameplusfam.net> wrote in message news:hqd6kc.4go.ln_at_mercury.downsfam.net...
>
> The objection to this seems to be that the
> RDM requires an implied unique constraint on all columns of all tables.
> Because such a constraint is not present, they are not truly relational.
>
> Is that right?

Yeah. I'm not sure about "implied" though.

I've never seen a relation in a math textbook where the uniqueness constraint wasn't on all attributes, but as we all know, there's lots of uses for keys with fewer than the maximum number of columns, and even for multiple keys. These are necessarily explicit.

Anyway, the definition of "relation" certainly includes a uniqueness requirement, no matter how you slice it.

> What are some of the other objections?

I think NULLs have been mentioned. I suddenly have the idea that a critique of NULLs is more properly a critique of the SQL type system rather than of its relational capabilities (although the two are certainly related.)

NULLs certainly suck in a lot of ways. In the context of the semi-mainstream definition of 1NF (what Alfredo would call the broken definition) NULLs are hard to get away from, because you need some operation like LEFT OUTER JOIN. If you have Relation Valued Attributes (RVAs) and some GROUP BY operator, though, you don't need NULLs even for OUTER JOINs.

Marshall Received on Sat Oct 09 2004 - 17:59:03 CEST