Aggregation
The selected key figures are aggregated in the Analytic Manager in the BW system. The aggregation behaviour determines how key figure values are aggregated in queries that use different characteristics and characteristic values.
In order to calculate the values of key figures, the data from the InfoProvider has to be aggregated to the detail level of the query. Formulas might also have to be calculated. A number of characteristics need to be aggregated here.
The basic form of aggregation is Standard Aggregation. When performing standard aggregation, the system can aggregate key figures according to the following rules: Summation (SUM), Minimum (MIN), and Maximum (MAX). The standard aggregation is independent from a specific characteristic. The system performs standard aggregation over all charactertistics that are not in the breakdown, irrespective of their sequence. This is the most well-known aggregation behavior.
In DataStore objects for direct update, it is also possible in planning mode to use key figures with standard aggregation NO2, no aggregation (X if more than one value appears).
In addition to standard aggregation, there is also Exception Aggregation. The aggregation rules for exception aggregation depend on a characteristic. You define how a restricted, calculated key figure or formula is aggregated in the Analytic Manager with regard to one or more reference characteristics. (Calculated key figures and formulas mostly behave accordingly; Only in exceptions can there be differences in the query results). A characteristic is needed for exception aggregation in order to define the granularity with which the aggregation rule is applied. The system aggregates the key figures to this detail level according to the rules of the standard aggregation. The exception aggregation is also always an extension to, and not a replacement for, the standard aggregation.
Let us take aggregation rule (AVG) as an example: The system calculates the average from the sum of all values divided by the number of entries. To define the number of entries, a granularity of data must be defined to which the system is to aggregate the multi-dimensional data according to the rules of standard aggregation. The result of the standard aggregation is the starting point of the exception aggregation, as the basis for the calculation according to the Average of All Values rule.
From a business point of view for example, exception aggregation arises if warehouse stock cannot be totalled up at the time, or if counters count the number of characteristics for a specific characteristic.
You can find an overview of the aggregation rules in the system under Rules for Standard and Exception Aggregations.
You set the various types of aggregation behavior of standard and exception aggregation for basis key figures in InfoObject maintenance (see Tab: Aggregation), for formulas and calculated key figures in BEx Query Designer (see Selection/Formula/Cell Properties).
You can also overwrite the settings for aggregation with the settings in the query. In the documentation for the general analysis functions in BEx, see Local Calculations, Calculate Results as..., and Calculate Single Values as... (local aggregation).