The level of detail in your star model is called grain. The granularity is directly related to the kind of questions you expect your model to answer. Let's see some examples.
Suppose that we build a fact table with sales along with the following dimensions: product type, manufacturer, region, and time. With this model, we would be able to answer the following questions:
- Beyond puzzles, which kind of product is the best sold?
- Do you sell more products manufactured by Ravensburger than products manufactured by Educa Jigsaws?
What if you want to know the names of the top ten products sold? You simply cannot, because that level of detail is not stored in the proposed model. The dimensions contain information ...
