Same Fact Table for Multiple Partitions
When the same fact table is used for more than one partition in a cube, it is important that the same rows are not used in more than one partition. It is possible for a row that is used in more than one partition to be included multiple times when the cube is processed or queried; this can cause queries to return incorrect data.
You can use filters in partitions to ensure that data is not duplicated among the partitions. A partition's filter specifies which data in the fact table is used in the partition. It is important that the filters for all partitions in a cube extract mutually exclusive datasets from the fact table. For example, these filters are mutually exclusive within each set:
Set 1:
"SaleYear" = 1997
"SaleYear" = 1998
Set 2:
"Continent" = 'NorthAmerica'
"Continent" = 'Europe'
"Continent" = 'SouthAmerica'
Set 3:
"Country" = 'USA'
"Country" = 'Mexico'
("Country" <> 'USA' AND "Country" <> 'Mexico')
When you create mutually exclusive filters for partitions, ensure that the combined partition data includes all data you want to include in the cube.