Professional SQL Server™ Analysis Services 2005 with MDX

7.3. Removing Empty Cells

The number of cells in a cube is the uniquely identifiable space within the cube. It is the product of the number of members in each attribute of each dimension. This is referred to as the cube space. As you can imagine, the entire cube space can be quite large. Of the entire cube space, the cells which constitute the product of attribute hierarchies of each dimension can potentially have fact data. However, in a typical cube most of these cells will not have fact data. For example, take a simple cube that contains dimensions product, time, and store. Assume the fact table contains IDs for dimensions, product time, and the sales amount. The product dimension table typically contains columns pertaining to the dimension, such as product name, product category, product weight, product color, and discount. The store dimension table would contain information about the store such as city, state, country, and number of employees. The time dimension might contain day, month, and year. As the owner of existing stores you might be interested in looking at the sales of various products in stores across various time periods every week, month, or quarter to make a decision on what product lines to enhance to grow your business. The store manager may be interested in identifying the sales of the products along with discounts so that he can stock products that sell the most while having discounts to maximize the profit of the store. Hence, the types of questions that ...