I am confused about how A/S handles aggregations on virtual
dimensions. We have a cube with a small number of physical dimensions
together with a large number of virtual dimensions. In most cases the
virtual dimension members are based on the same exact key fields as
the underlying dimension (see example below).

Example:

Physical Dimension: Product by Category
Levels: Category, Department, Item
At each level, the member is identified by a key that is unique
throughout the dimension and a separate Name field which is not
unique.

Each of these levels also has a member property "Spanish Name"

Virtual Dimension: Product by Category Spanish
has the exact same levels and keys as Product by Category, but with
name based on the "Spanish Name" property.

My questions:

1. Why when I add a virtual dimension to my sales cube do I need to
recalculate aggregations? I would have expected the aggregations
based on the existing underlying physical dimensions would be used.

2. In a situation like this is it best to use Top-level aggregations
only?


Thanks for clarifying this.