Re: Relational Databases and Their Guts

From: Todd Bandrowsky <anakin_at_unitedsoftworks.com>
Date: 22 Jun 2003 16:41:45 -0700
Message-ID: <af3d9224.0306221541.211f9cae_at_posting.google.com>


> But SQL has union and intersection directly in select statements. It also
> has "except". What about "except" is not subtract that you are expecting?
> Or were you not aware of "except?" (It is somewhat obscure, and I don't
> believe it's present in SQL Server.)

SQL Server has NOT EXISTS, which could be shoehorned into this kind of work, yes. I believe Oracle also has the same. Many other database systems do not.

> Note also that this is a model-level description, and so the question of
> which one is "efficient to implement" is specifically excluded; the
> implementation may be anything that can represent the model.

agreed.

> Au contraire; this is where SQL stands tall, by not allowing such a
> thing.

But what if I needed it? Therefor, I have to do it in some other system, so, I can't have my database be a single repository for handling and manipulating all information, and, for that reason, I can't have any system that satisfies Logical Data Independence, unless it is through some middle tier of code, and not the database. Received on Mon Jun 23 2003 - 01:41:45 CEST

Original text of this message