Re: Normalization by Composing, not just Decomposing
Date: Mon, 12 Apr 2004 16:18:32 GMT
"Alan" <alan_at_erols.com> wrote in message
> Realtional theory as a data theory is analagous to democracy as a form of
> government- it may not be perfect, but so far, there's nothing better in
> most cases.
> Denormalization in itself has nothing directly to do with OLAP, except
> one may denormalize more for an OLAP application than an OLTP application.
> However, in OLAP, you are not necessarily denormalizing so much as
> "re-normalizing", in that you are really developing a diiferent
> among entities for the same data, such as in a star schema. It's not
> normalized, but it's not denormalized either. It's just different.
That's just nonsense. At least have the courtesy to define new terms - "normalize" was created and defined fairly precisely by the relational camp. Star schemas duplicate information in the alleged name of performance, and as such are less normalized, not just "different."
> I suppose
> an argument could be made that (in the case of a star schema), you start
> with a normalized schema, and then apply transformation rules (no, don't
> me what they are- there are books on the topic) to transform it into a
> schema. Think about it- a basic star schema is essentially a giant
> many-to-many linking table (the fact table) with a bunch of descriptive
> tables (dimensions).
It's far more than that. The fact table typically combines many different predicates, again in the name of performance (to avoid joins). So of course it's not normalized. It's normalized if you squint really hard and bang your head against the table a few times, then look at it.