Handling User-Defined Data
Roy Hann, first presented at UK IUA Autumn Conference, Market Bosworth 2005

 

Home

We frequently encounter the need to extend the database design and applications to include so-called "user-defined data". Changes and improvements to business processes depend on having some additional details that were not anticipated in the original design. Sometimes entire new fact types must be added. Or sometimes truly massive new extensions to the database must be allowed on an almost ad hoc basis. In this paper I describe some of the details of these problems, and one frequently proposed solution (the entity-attribute-value design, or EAV). I then describe a much simpler, more flexible, and powerful solution the exploits the intrinsic capabilities of a relational DBMS like Ingres.

The presentation is in the form of 27 Microsoft PowerPoint 2000 slides, zipped to 101kb and inflating to 212kb.

Topics

>> What "User-Defined" data is
>>
The Entity-Attribute-Value design
>>
A few of the many problems with EAV
>>
A better way
>>
So the DBMS can handle it; so what?
>>
Using dynamic SQL and system catalogues
>>
User-defined data is exactly what the relational model is for

Download now. (101kb)


Ingres® is a trademark of Ingres Corporation | ©Rational Commerce Ltd. | Last updated 13/11/2006 11:39