Database Reference
In-Depth Information
ant difference during dimension processing; you will learn more about processing in
Chapter 5 , Optimizing Dimension and Cube Processing .
User hierarchies do not always have to include relationships between each combin-
ation of levels. For example, users might want to browse data by the promotion start
date and the discount percent. However, the same discount percent might be ap-
plicable to multiple start and end dates, depending on the quantity of the product
a customer decides to purchase. Therefore, we have a many-to-many relationship
between the start date and the discount percent, and we cannot define an attribute
relationship between these two levels. If a hierarchy contains attribute relationships
defined between each combination of levels, it is called a natural hierarchy. Analys-
is Services builds additional indexes for natural hierarchies, which help optimize the
query performance. Therefore, to achieve optimal query performance, always define
the necessary attribute relationships when applicable.
All other attributes included in the Promotion dimension (minimum and maximum
quantity, start and end dates, discount percent) will be directly related to the pro-
motion key because each may be applicable to multiple promotions. Once you've
defined the relationship between the promotion type and category attributes, save
Search WWH ::




Custom Search