Dimensional explosion
As soon as we introduce aggregations, the data in the cube is partitioned into two categories: raw data and derived data. Combinatorial exercises show that the number of the combinations of aggregated members in each dimension with all other members is huge. This is referred to as dimensional explosion.
How big is it? We often hear users say: “Well, in my dimension hierarchies 90% of members correspond to the input level. I think the amount of aggregated values is not significant compared to the raw data.”
A simple computation shows that 0.97»0.478. Hence with 7 dimensions more than 50% of the values are derived! And with 10 dimensions over 65% of the values are derived!