Creating Analysis Services Cubes - SQL Server Analysis Services 2012 Cube Development - page 112

</Source>
</Sources>
<Target>
<DatabaseID>ssas_cookbook_chapter3</DatabaseID>
<CubeID>AdventureWorksSample</CubeID>
<MeasureGroupID>Fact Reseller
Sales</MeasureGroupID>
<PartitionID>Fact Reseller
Sales</PartitionID>
</Target>
</MergePartitions>
Unfortunately merging partitions does have its share of limitations. For example, the
query definition of the resulting partition will be the same as the query definition of
the target partition. Let's suppose we had set the target partition's query to only in-
clude 2005 sales. Once we merge 2006 and 2007 into the target, the resulting parti-
tion will contain data for all three years. However, the query definition will remain set
to only include 2005 data. Now, if you intentionally or accidentally process the result-
ing partition, you will lose data for 2006 and 2007. Fortunately, you can update the
partition's query definition without having to reprocess it.
Another limitation is that MergePartitions does not update the target partition's
slice. So the resulting partition could actually provide a wrong query hint to the serv-
er, indicating that it only contains 2005 sales whereas in reality it contains sales for
2005, 2006, and 2007. To work around this limitation, you could create a new parti-
tion with the correct slice, include a query definition that returns no rows (" SELECT
column1, column2 … FROM fact_table WHERE 1=2 "), process the empty
partition, and then use it as the merge operation's target.