Understanding dimension processing options

We must ensure that each dimension referenced, by a particular measure group, has an up-to-date dataset prior to processing any partitions in that measure group. If we attempt to process a measure group before dimensions have up-to-date data, we will encounter missing key errors and the measure group / partition processing will fail.

How it works...

The following table summarizes dimension processing options:

Processing option	Description	Used for
ProcessFull	This option loads a full dimension dataset and rebuilds all indexes. This option invalidates all partitions referencing the current dimension; every partition will have to be fully reprocessed.	Each dimension must first be processed using ProcessFull. From time to time, you should reprocess these dimensions to remove fragmentation introduced by ProcessAdd or ProcessUpdate.
ProcessClear	This option empties a dimension object's contents to "unprocess" the dimension.	...