Re: Question on Structuring Product Attributes

On Tue, 18 Oct 2011 19:35:12 -0700 (PDT), -CELKO- <jcelko212_at_earthlink.net> wrote:

>>> As Joe can confirm, I often disagree with him, on many subjects. But when I first saw him post the subtype/supertype implementation model you criticise, I liked it for its robust protection against integrity violations. You are correct that the redundant columns and the extraneous keys require more disk space and hurt performance, but I always think that integrity comes first, all the rest comes later. <<
>
>Hey, we do not disagree very often, where dis you get a silly idea
>like that? Sorry, I had to do a Monty Python arguer skit :)

Hi Joe,

Back in the days when the SQL Server usenet groups were still lively (before Microsoft replaced them with something worse), I often challenged you when you posted suggestions that included code that does not work well on SQL Server, or code that would make SQL Server grind to a halt because of its awful performance. But I will admit that Tony Rogerson often beat me to the punch.

(...)

>Wait until he sees Kuznetsov's pattern for preventing temporal gaps in
>event sequences or my state change constraints :) Overlapping and
>self-referencing uniqueness constraints are hard for an experienced
>SQL programmer.

Have you read Kuznetsov's pattern in his Defensive Database Programming book, or on the web? In the former case, you might want to check out who tech edited it, just for laughs. ;)

-- 
Hugo Kornelis, SQL Server MVP
My SQL Server blog: http://sqlblog.com/blogs/hugo_kornelis