I am probably just missing something here, but I've read a couple places now that SSAS does not store aggregates except for the leaf level and the All level of parent-child hierarchies and that they should be avoided if possible. I am wondering why this
is? I guess I just don't see the difference between storing fact summaries for members of a parent child hierarchy vs. storing them for attributes with configured hierarchical relationships in a non p-c dimension. I am writing an application where data is
modeled as belonging to a member of each of several ragged hierarchical dimensions that are configured with the parent-child recursive relationship. I would like to use a cube for performance purposes on the reporting end, but now I am having to flatten these
dimensions and use the HideMemberIf property to achieve what I need. It feels like I have to fool SSAS into doing what it very easily does otherwise. Anyways, just curious as to why it's a problem for SSAS to store the aggregates for a parent-child hierarchy
when with a bit of extra configuration you can achieve the same / similar effect.
↧