
Silverston c09.tex V2 - 11/21/2008 3:16am Page 513
Data Warehouse Models 513
In conclusion, the data team created a very flexible data model that attempted
to support the future needs of the business and maintain a history of changes
in relationships. In fact, they also decided in this data warehouse data model
to use the same level of generalization throughout the model, namely, level 3
patterns. In doing this, they believed that they would give up some simplicity
of implementation, but in practice, because the models were very consistent,
there was a great benefit in being able to consistently manage the data in
the data warehouse and reuse routines. For instance, they noticed that the
part of the data model that supported flexible classification of customer
information was the same pattern that supported product classification (or
any other classification). This meant that the routines and code supporting
product classifications could be reused, with some modification, to support
any classification. They also noticed that some of the constructs that were
reused, such as the status pattern and the contextual role pattern, had useful
code and expertise surrounding them in the customer application that they
could leverage. Finally, developing the data warehouse model based on the
enterprise data model is a well known ‘best practice’ and seemed quite natural
and intuitive. Some of the data team was involved ...