Measures and Multidimensional Analysis - Microsoft SQL Server 2008 Analysis Services - page 80

</Measures>
<Dimensions />
<Partitions />
</MeasureGroup>
This example is not a full example. Some properties are missing from the DDL, such as the
definitions of dimensions (to be explained later) and partitions (part of the physical
model). In addition, certain properties, such as AggregateFunction , have default values
assigned when they are missing from the DDL.
Measure Group Dimensions
A measure group dimension is a list of the cube dimensions that belong to the fact (measure
group). The data space (dimensionality) for a fact is a subset of data space of the cube,
defined by the measure group dimensions (that are a subset of the cube dimensions). If
the list of dimensions in the cube is the same as the list of those in the measure group,
you could say that the measure group has the same dimensionality as the cube. In that
case, all the dimensions in the cube define how the data is loaded into the measure group.
However, most cubes have multiple measure groups, and each measure group has its own
list of dimensions.
In our example, the Warehouse measure group contains only four cube dimensions:
Product , Time , Store , and Warehouse . All the other cube dimensions are unrelated to the
Warehouse measure group (or to the measures of the measure group). Unrelated dimen-
sions don't define how the data is loaded into the fact. When users request data using an
unrelated dimension in a query, Analysis Services based on the IgnoreUnrelatedDimensions
measure group property either returns a NULL value or uses other measure group dimen-
sions to resolve the request. For example, suppose a user requests the value of Units
Ordered measure by specifying coordinates on the Time , Product , and Customer dimen-
sions. Whereas Time and Product dimensions are related to the Warehouse measure group,
the Customer dimension is not related.
A measure group dimension doesn't have its own name or ID. A measure group dimension
is referenced by the ID of the cube dimension. The Annotations property is the only prop-
erty that a measure group dimension has.
Granularity of a Fact
Each dimension of a measure group defines the rules for loading the fact data for that
dimension into that measure group. By default, the system loads data according to the key
attribute of the dimension. In our sample, the day defines the precision of the data for our
Sales measure group. In this case, we would say that the Date attribute of the Time
dimension is the one that defines the granularity (or the depth) of data for the fact.
The attributes that define granularity in each dimension of the measure group, taken
together, make up a list called the measure group granularity . In our sample, the measure
group granularity of the Warehouse measure group is a list that contains the Product ,