Re: 1 NF

From: dawn <>
Date: 28 Feb 2007 12:56:52 -0800
Message-ID: <>

On Feb 28, 1:14 pm, "frebe" <> wrote:
> Hi,
> I read an interview with Chris Date (
> int9410.html) which made me a little bit confused. He claims that
> having an array as a column values doesn't violates 1NF. Is this the
> common opinion at comp.databases.theory too? If yes, how is it
> possible at apply 1NF at all, in that case?
It seems that Date and other relational theorists have decided to redefine "1NF" in recent years so that it means almost nothing. So, whenever I want to mention 1NF here, I use something like "the form formerly-known-as-1NF." This becomes all the more confusing when working with NF2 data, which stands for "Non-first Normal Form." This is non-1NF by the former and not more currently accepted definition. Ugh!

Data modelers who work with SQL-DBMS's are still modeling with the form-formerly-known-as-1NF approach. Practice typically follows theory, so that might be expected. But even on the theory side, I have not seen any good writeups to help data modelers who work with NF2 DBMS's or who could use Arrays know when to model data with such an approach (although those who have been working with NF2 databases for decades certainly have established some best practices).

Cheers! --dawn Received on Wed Feb 28 2007 - 21:56:52 CET

Original text of this message