I have a measure group with both conventionally aggregated measures (aggregation type SUM) and semi additive measures (LastNonEmpty). The cube processes them quite happily, and I can view and analyse (slice) them fine, but when I analyse them both together, then the conventionally aggregated measures (aggregation type SUM) always dissapears in favour of the semi additive measures. This is not the case if I analyse conventionally aggregated measures (aggregation type SUM) and semi additive measures (LastNonEmpty) from different measure groups.Why does this happen, and is there a way to overcome this by pref not creating seperate fact sources for the conventionally aggregated measures (aggregation type SUM) and semi additive measures (LastNonEmpty) currently in the same fact source.

**View Complete Post**

## MDX - ParallelPeriod using a semi additive measure