Data Warehouse Designs

10

Why Is Time Variance Difficult?

In my previous book, Building and Maintaining a Data Warehouse, I demonstrated how to build time variance into a data warehouse. In that discussion, an object in a row of a fact table joins to a dimension table via a primary key/foreign key join with an additional WHERE clause that selects the one row of the dimension table that has the Effective and Termination Dates that form a time interval that surrounds the date of the transaction in the fact table row (Silvers 2008). This and subsequent chapters constitute both a retraction and a correction of that time variant design. In practice, I have found that this model performs well at low cardinalities of data. I found that as long as the number of fact table ...