I am wondering what the best practices are regarding an abstract vs. a
specific database design in regard to OLAP. Here is what I am doing:

I have a "Person" table.
I can create an abstract set of tables to hold Attributes of a Person, such
as Gender, Ethnicity, etc, called "Attribute Type" (To hold Gender) and then
a table called "Attribute Value" (to hold Male), and put the PersonID in the
"Attribute Value" table.

OR

I can create a "Gender" table, and put the GenderID in the Person table.

My question is this:
What are the performance and functionality losses in regard to creating OLAP
cubes, that come from going with the first, more abstract way of modeling a
database?

tx

I didn't mention the obvious connector table between Attribute Value, and
Person that will hold the AttributeValueID and the PersonID.

"Tim Chapla" <tchapla (AT) nospam (DOT) com> wrote