Optimizing Dimension and Cube Processing - SQL Server Analysis Services 2012 Cube Development - page 181

2. Much like what you saw with dimensions, SSDT allows choosing multiple
partitions if you wish to process them in the same batch. To use SSDT for
partition processing, navigate to the Partitions tab within the cube design-
er, expand the measure group you're interested in, and click on the partitions
you want to process while holding down the Shift key.
3. Both SSDT and SSMS allow for setting the same options as with dimension
processing. The options you could specify are the degree of parallelism, di-
mension key errors, and whether the entire batch should be processed in a
single transaction.
Dimension errors could occur during the partition processing if the partition
query (or table) references a dimension key that does not exist in the SSAS
dimension. This is the reason why you should process dimensions prior to
processing any partitions.
Tip
Exercise special care when processing many partitions in parallel. SSAS
sometimes attempts to bite off more than it can handle unless you expli-
citly specify the number of partitions to be processed in parallel. Further-
more, the relational source could also get overburdened if you attempt
processing too many partitions in parallel, because processing each one
requires sending a hefty query to the relational database.
4. Sadly, SSDT does not provide an option to script the processing command.
Not to worry, you can get the XMLA command from SQL Server Profiler . For
example, the following command processes three partitions in parallel:
<Batch xmlns =
"http://schemas.microsoft.com/
analysisservices/2003/engine">
<Parallel MaxParallel="3">
<Process>
<Object>
<DatabaseID>AdventureWorks_Sample</DatabaseID>