Database Reference
In-Depth Information
Unlike UBO and
Aggregation Design
wizards,
Aggregation Manager
does not lim-
it the size of aggregations, nor do you have to provide row counts for each attribute.
Clearly, you should be careful not to design too many aggregations,so that you don't
run out of disk space or tremendously increase the processing time. However, I re-
commend becoming familiar with
Aggregation Manager
in case you need greater
control over aggregation designs than that offered by SSDT/SSMS wizards.
There's more...
As mentioned earlier in this chapter, you can use the
ProcessIndexes
processing
option to build indexes and aggregations on a given partition for which data has
already been loaded (The
ProcessFull
option includes
ProcessIndexes
). In
some environments it might be beneficial to make data available to users even be-
fore indexes and aggregations are built. MOLAP storage often outperforms relation-
al queries by an order of magnitude, even if you don't build any aggregations and
indexes, thus business users can query data faster even before we add the speed-
boost with indexes and aggregations. In this case you can allow Analysis Services to
Search WWH ::
Custom Search