64 IBM Cognos Dynamic Cubes
A member could exist in multiple locations in a level but each instance would be qualified by
the level unique key so that the context that makes each instance unique would be identified.
Measure dimensions
Measure dimensions contain the fact data, or measures, that you will use in your reporting
and analysis. A measure dimension is based on a single fact table in the data source. A
dynamic cube can have any number of regular dimensions, but only a single measure
dimension based on a single fact table in the data source. If your application requires using
measures from multiple fact tables, see Chapter 6, “Virtual cubes” on page 133 to learn more
about virtual cubes.
4.2.4 Creating calculated measures
You can extend the usefulness of cubes by creating calculated measures, which have custom
expressions that your consumers will want to use. The expressions can use the rich function
library, available to create calculated measures. With the calculated measures, you can build,
into your cube, measures that are complex and take considerable time to generate results. By
building them into the cube, they need to be created only once. It is not necessary to
re-create the expression for each report that your consumers need to use. The administration
of each measure is easier, too. Because the measure is built into the cube, the performance
of reports that use it will be faster.
Report objects
Cognos reporting applications use two types of metadata in their packages that are available
to report authors. They provide unique features that can help you fulfill your reporting
requirements.
The first type is the metadata that you create when you model in Cognos Cube Designer.
These are the dimensions, hierarchies, levels, attributes, and measures. For example, using a
level in a report will always get the members that belong to that level. The report does not
need to be updated as the dimension changed.
Chapter 4. Modeling dynamic cubes 65
Figure 4-3 shows the metadata that is created when you model in Cognos Cube Designer.
Figure 4-3 Creating metadata
The second type is the members that are generated in the cube as a result of modeling, as
shown in Figure 4-4.
Figure 4-4 Members generated in the cube
How you want to represent and model your dimensions is your choice. A regular dimension
can have many hierarchies. Using multiple hierarchies within a dimension may help organize
similar hierarchies. Depending on what your dimension is like, you could have many possible
hierarchies.
There is no effect of the presence of a hierarchy on the other hierarchies in the dimension.
The query that generates the members for each hierarchy is separate. If, however, objects
from multiple hierarchies are included in a report, the tuple combinations that do not exist in
the fact will be removed from the set during query processing. The effect in the report is to
filter it.
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.