Executing the Aggregation

The OLAP DML provides two ways to aggregate data:

• As a data maintenance procedure using the AGGREGATE command. If you want to use this method of aggregating data, then, within an aggregation specification, identify data that you want to aggregate in this manner using the PRECOMPUTE statement or PRECOMPUTE clause of the RELATION (for aggregation) statement.

• At run-time when needed using the AGGREGATE function or adding an $AGGMAP property to the variable.

You can choose whatever method seems appropriate: by level, individual member, member attribute, time range, data value, or other criteria. You can also combine these approaches and precalculate some values and calculate others at run time. In this case, frequently, you use the same aggmap with the AGGREGATE command and the AGGREGATE function. However, in some cases you might use different aggmaps.

One step that you can take to achieve overall good performance is to balance the amount of the data that you aggregate and store in an analytic workspace with the amount of data that you specify for calculation on the fly. You can use a PRECOMPUTE statement or clause within your aggregation specification to ask Oracle OLAP to use special functionality called the Aggregate Advisor to automatically determine what values to aggregate as a data maintenance procedure using the AGGREGATE command, or to explicitly identify the values yourself.