Database Reference
In-Depth Information
One thing to remember is that typically, a single partition is all you really need. If at some point you realize
that you have defined more partitions than you need, you can delete a partition by right-clicking the partition and
selecting Delete from the context menu (Figure
12-15
).
Figure 12-15.
Deleting an existing partition
■
deleting the partition does not mean the data the partition used to contain will be magically added to
the other partitions. You must change the source description or sQl code on any existing partitions to make sure you
are obtaining all the data. For example, we could delete both partitions defined in Figure
12-14
, but we would then
have to make a new partition that covers all years by selecting the entire Fact sales table as the source.
Caution
Partition Storage Designs
Earlier, in Figure
12-14
, you may have noticed two hyperlinks on the Partitions tab. The first hyperlink allows you
to launch the Partition Wizard, but the second hyperlink opens the Partition Storage Setting dialog window.
In this dialog window you can choose how data is stored for a particular partition. Partition data falls into
three categories of data: leaf level (individual values), aggregated total and subtotal values, and metadata that
describes the partition.
Data storage designs also fall into three categories: MOLAP, HOLAP, and ROLAP (Figure
12-16
). Storage
locations, however, fall into two categories: an SSAS database folder or tables in a relational database.
