Microsoft Office Tutorials and References
In Depth Information
In this chapter, you learned several useful concepts about data models.
■ A data model is nothing but a set of tables and relationships that describe some entities.
Relationships are the most important part of a data model, giving entities a meaningful
position in the data model.
■ The same entity structure can be described by different data models, each one
having virtues and laws. The choice of the right data model seriously affects the
way we query it.
■ PowerPivot is a database, so it is able to understand and use data models. This leads to
the necessity, for the PowerPivot user, to understand and correctly use data models.
■ The physical data model is often complex because of the normalization process that
has been carried on by database administrators to optimize software that runs on it.
On the other hand, the logical data model needs to be as simple as possible because
it will be queried by humans. The basic process you learned to convert a physical data
model into a logical one is called denormalization.
■ There are a couple of ways to denormalize data: in PowerPivot or through SQL queries.
The PowerPivot way is easier but less powerful; the SQL one is much more powerful,
but it requires of you some basic SQL understanding.
■ If we have a data warehouse, querying it is the best choice because data in the data
warehouse is already denormalized and probably stored as facts and dimensions, a
very simple structure to query that provides great results.
■ Even a structure as simple as the data warehouse often contains complexities, such as
many-to-many relationships or role-playing dimensions. Those advanced dimensions
need a specific treatment and understanding.