Deploying and Processing Cubes - Getting Started with SQL Server 2012 Cube Development - page 96

[dbo_FactInternetSalesShipDateKey0_3],[DueDateKey]
AS
[dbo_FactInternetSalesDueDateKey0_4],[ProductKey]
AS
[dbo_FactInternetSalesProductKey0_5],[CustomerKey]
AS
[dbo_FactInternetSalesCustomerKey0_6],[PromotionKey]
AS
[dbo_FactInternetSalesPromotionKey0_7],[SalesTerritoryKey]
AS [dbo_FactInternetSalesSalesTerritoryKey0_8]
FROM [dbo].[FactInternetSales]
)
AS [dbo_FactInternetSales]

Processing actually reads everything from the source. This can be seen in the pre-
ceding code as it just runs a select of all columns without the where clause. This
can be very costly, later in this chapter, we will look at ways to limit this burden for
the relational database through the use of partitions.

Using advanced processing options when

processing cubes

There are also advanced processing settings that you can specify. We earlier
covered parallelism which is one of the settings, another one is how it should handle
transactions. By default, Analysis Services tries to execute processing in parallel but
as one transaction. You also have the possibility to configure it to execute in sequen-
tial mode. This means that it will process every object sequentially when you config-
ure processing. In this way, you also have the possibility to configure if it is be done
in separate transactions.

When you process an object related objects also need to be processed. As an ex-
ample, if you process a dimension, you have to process related cubes and measure
groups. You can determine what objects will be processed by clicking on the Impact
Analysis... button in the process dialog.

The other options around error handling will be covered later in the chapter.