Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!iuvax!cica!tut.cis.ohio-state.edu!cs.utexas.edu!uunet!mcsun!hp4nl!eurtrx!euraiv1!reino
From: reino@cs.eur.nl (Reino de Boer)
Newsgroups: comp.databases
Subject: Unique code numbers
Message-ID: <1990Jan16.081519.279@cs.eur.nl>
Date: 16 Jan 90 08:15:19 GMT
Organization: Erasmus Universiteit Rotterdam, dept. CS (Informatica)
Lines: 38

For the first time in my life I am planning a reasonably large database,
and I have run across a (solvable) problem.

A lot of examples in literature about relational databases involve the
following kind of relationships:

Supplier = (Name, Address, etc., Parts supplied)
Part = (Name, Description, etc., Suppliers)

which is then normalized to

Supplier = (SupplierID, Name, Address, etc.)
Part = (PartID, Name, Description, etc.)
SP = (SupplierID, PartID)

I was wondering if there is a relational DBMS that gives me the
opportunity to indicate that:
1. SupplierIDs need to be unique
2. PartIDs need to be unique
and, most important of all
3. Frees me from the burden to make up new SupplierIDs and PartIDs.

The problem is solvable by introducing a new relation for each ID:
IDconfig = (lastID)
and generating the new unique ID by something along
ID <- retrieve( lastID )
update( lastID )
store( lastID )

but I would like a more general mechanism. Perhaps a special field data
type `unique id', that generates a new value when needed.

Awaiting your replies, thanks already -- Reino

-- 
Reino R. A. de Boer
Erasmus University Rotterdam ( Informatica )
e-mail: reino@cs.eur.nl