Re: Attribute-value approach for table design
- From: "Jamie Collins" <jamiecollins@xxxxxxxxxx>
- Date: 13 Nov 2006 00:35:56 -0800
Jeff Boyce wrote:
Can you offer some pointers to arguments both against and for EAV? A
cursory Google search suggested that there are performance issues in this
design, but that for something like a medical tests database, EAV offers a
practical solution.
The intent of my post was to give a name to the design, using the word
'flaw' in quotes to reflect the low opinion in general application, to
encourage the OP to do some research. Mission accomplished ;-)
Let's face it: the chances of EAV being a newbie error, rather than
considered design choice, are extremely high.
Bear in mind that 'database' and 'SQL DBMS' are different concepts. My
impression from reading around the subject is that if EAV is the
required solution then SQL is not the best means to achieve those ends
i.e. 'performance issues' are a symptom of a fundamental bad fit.
My feeling is that the limited applications that benefit from EAV would
not be suited to an Access/Jet SQL.
Any additional evidence, or are we in the realm of opinion?
I've no first hand experience to add.
Jamie.
--
.
- Follow-Ups:
- Re: Attribute-value approach for table design
- From: Jeff Boyce
- Re: Attribute-value approach for table design
- References:
- Re: Attribute-value approach for table design
- From: Jamie Collins
- Re: Attribute-value approach for table design
- From: Jeff Boyce
- Re: Attribute-value approach for table design
- Prev by Date: Re: An array within a table--Break into 1-to-1 mini tables or separate rows?
- Next by Date: Re: How do I make the Primary Key autonumber as date and time
- Previous by thread: Re: Attribute-value approach for table design
- Next by thread: Re: Attribute-value approach for table design
- Index(es):
Relevant Pages
|