Re: Looking for a discussion about generic datamodels

From: Kenneth Downs <knode.wants.this_at_see.sigblock>
Date: Fri, 02 Sep 2005 09:41:43 -0400
Message-Id: <8c1mu2-7mi.ln1_at_pluto.downsfam.net>


schreurs_roel_at_hotmail.com wrote:

> Every now and then, I come across a IT-project that stores its data in
> a generic data model. Such data models basically consist of 4 tables:
> Entities, Relations, Attributes and Values. The Entities table contains
> a record for each table in the conceptual data model, Attributes a
> record for each column, etc.
> Invariably, the choice for such a data model is defended by the
> argumentation that new conceptual tables and columns can be added
> without modification of the data model.
> This all sounds very nice, but I always have the idea that this comes
> at a price that may well be higher then the benefits it offers.
>
> So, what I am looking for is a discussion, be it on line or in
> literature, about the pro and cons, as well as some list of
> recommendations to help me decide if such a model would be useful. Can
> anybody recommend a source of information?
>
> Roel Schreurs
> The Netherlands

To expand on what Roy said about queries, when a person makes this design blunder they end up having to reinvent SQL. SQL was crafted to read and write columns from tables. The E-A-V blunder transposes columns to rows, now requiring a transpose step for every single column of every single query. I don't even want to think about joins, group bys(!!). Ouch.

Much cheaper is to attack the real problem: the cost of changing table structures and keeping code synchronized.

-- 
Kenneth Downs
Secure Data Software, Inc.
(Ken)nneth_at_(Sec)ure(Dat)a(.com)
Received on Fri Sep 02 2005 - 15:41:43 CEST

Original text of this message