Aggregation is a simple aggregation based on the Cube Dimension hierarchy. Additional aggregation points for non-cube / attribute members can be included in the results using Derivative Transformation Rules.
The BI Blend Engine aggregation utilizes the Cube Dimension hierarchies. This is done by evaluating parent members and identifying all Base members within the hierarchy. These base members are aggregated to the parent. Each parent within the hierarchy is evaluated using the same methodology. BI Blend processing has no concept of “sub-parent” rollups. Each parent is evaluated and aggregated according to the base members within its hierarchy.
Entity properties such as “Percent Consolidate” and other Dimension’s “Aggregation Weight” are not used in the BI Blend processing. The aggregation is derived strictly from each Parent as a sum of its base members. Therefore, duplicate members within a hierarchy (shared members), should be avoided to eliminate “double-counting” of results.
BI Blend supports changing the “Blend Unit” from Entity as the “page dimension”, to any other dimension in the Cube. This requires that Currency translation to be defined as a simple translation based on the Entity's local currency. This is the case whether Entity is set as the Blend Unit or a different dimension is used. This allows the correct results when the Entity is not the Blend Unit, by providing a common currency throughout the aggregation levels to yield correct results.
Each defined aggregation is stored in a cache, and each can be calculated independently.
