Re: Re-design of data storage table

From: Palooka <>
Date: Fri, 27 Feb 2009 00:10:14 +0000
Message-ID: <GbGpl.31869$Dz4.12577_at_newsfe20.ams2>

Erland Sommarskog wrote:
> ( writes:
>> A few days of trawling the net is how you decide if your database
>> design is correct?
>> You are complaining that no one else can suggest a better solution
>> without paying for consulting by competent database designers?
>> A bunch of us experienced people are noting that this won't scale more
>> as a courtesy than anything else.  You get what you pay for eh?
>> Good luck but keep your resume updated and you may not want this
>> project on it.

> Maybe you don't you this post on your résumé?
> If you want to play a database designer, it does help to read want people
> want.
> Paul wants a database where users are able to add their own attributes, and
> there should be no limits. Notice that Paul is not a new upstart, but he
> has an existing system, which uses a different model than EAV, and which
> has the drawback of setting a limits the number attributes there can be.
> We don't know too many details about Paul's system, but apparently this
> limitation gives his product a drawback on the market.
> Yes, EAV is problematic, but the whole area of customer-defined attributes
> is problematic. Nevertheless, this is a common requirement in many
> systems, and the not least if you want to make a product of your
> application. EAV is probably the best solution for this situation. The
> fact that EAV can be abused is no reason to throw out the baby with
> the bathtub.
> I have used EAV in a few corners in the system I work with, and I can
> tell you: it has been a big life-saver.

A few customer-defined attributes on top of a reasonably defined model, yes.

Basis of the system? Recipe for disaster, as others have observed.

Palooka Received on Thu Feb 26 2009 - 18:10:14 CST

Original text of this message