Aggregation can be of the following type:
Table Level: Commonly known as Marts.L1/L2 marts at monthly/yearly levels.
Query Level: Make use of Database aggregate functions like SUM COUNT MAX
MIN to reduce number of dimensions and thus reducing number of rows returned by
the query.
Projection Level: Once BO has data in the micro cube we can assign projection
level aggregation. This is useful in Drill down functionality.
E.g A table contains name of the school class and names of all the
students in the class. want to count number of students in the class hence at
query level the measure will be defined as COUNT(Student_name). Hence the
resultset in microcube of BO will be like:
School Class Count
X A 10
X B 20
But I want to show number of students in school X. which will be 10+20. Hence
projection level agg will SUM and not COUNT. Projection level agg is set at
object level for measure in properties tab.