Creating a partition (as explained in the Creating and managing partitions recipe) creates a definition of the partition (in model metadata). Once this is done, the data must be loaded into the partition if it has to become viewable. Furthermore, changes to the underlying data are not reflected in the model until the partition is processed.
Consider the situation where data is partitioned at the source database. This could be by table, partitions within tables, or views (or various combinations of these). For example, assume that we have the sales data separated in views by year, as shown in the following screenshot:
The partitioning strategy of the model also follows this structure because sales are recorded on the date that the order is placed (assume that this is a business rule) and, therefore, only the values in
dbo.reseller_sales_2010 are expected to change in the source data when the data is updated (in a nightly batch). The least resource-intensive operation to...