Dimensional modeling When ER models get too complicated to query, dimensional modeling can offer a practical solution. A dimensional data model is composed of a single fact table. This fact table contains a compound primary key, with separate keys linking the fact table to the dimension tables. These dimension tables contain descriptions and attributes that provide context to the metrics stored in the fact table. Dimensions often contain data on multiple hierarchical levels that are "flattened" (or denormalized) into a single table. For example, in our Aircraft Types dimension table, we have both Aircraft and Aircraft Manufacturer in the same table.