Two types of aggregate tables are discussed.
The first is where all non-additive facts and one or more dimensions are removed from a fact table. Typically this results in a smaller table that can answer a subset of the queries that could be posed against the fact table. This aggregate table still maintains full integrity to the remaining dimensions, and consequently reflects all changes to those dimensions.
The second type, we will call an aggregate summary, or summary table. This table includes additive measures and in some cases hierarchical elements of one or more of the dimensions providing a rolled-up summary of the fact table data. For example we may choose to deal at product group level rather than product SKU which is the granularity of the dimension.