Re: taxonomy for constraints?

From: Tony Andrews <andrewst_at_onetel.com>
Date: 5 Sep 2005 07:32:35 -0700
Message-ID: <1125930755.237959.278870_at_o13g2000cwo.googlegroups.com>


Marshall Spight wrote:
> Hi all,
>
> Does anyone know of any interesting constraint taxonomies?
>
> I am aware of
> 1) D&D's domain/column/table/database hierarchy, which is
> intuitively appealing but I'm not convinced it has much behind it.

I know what you mean. A domain constraint is just a domain definition; a column constraint is just a column definition (in terms of a domain); and table constraints are just a special case of database constraints, i.e. those involving just a single table. It's not clear what the taxonomy "buys you".

> There are other kinds of things in the relational world, as well.
>
> Foreign keys
> uniqueness constraints
> functional dependencies

Foreign keys and uniqueness constraints are really just shorthands for commonly used constraints that could be defined more generically in the same way as any other constraint. There are other fairly common constraints that are not so distinguished, e.g. "Employee_Project_Assignment date range must fall within Project date range", "Employee salary must fall within the MimimumSal and MaximumSal for the Employee's Grade". These seem to be "like" foreign keys in some way, and yet clearly not the same.

> Anyone care to discuss how these all fit together? Anyone
> know of any formal treatment of these?

I'm sure I have seen some detailed taxonomies of constraints and rules on www.brcommunity.com, but can't find them right now. Received on Mon Sep 05 2005 - 16:32:35 CEST

Original text of this message