Logical Data Warehouse Design - Data Warehouse Systems: Design and Implementation - page 154

the Time dimension are given in Fig. 5.26 . In Analysis Services, there are two
types of relationships, flexible and rigid. Flexible relationships can evolve
across time (e.g., a product can be assigned to a new category), while rigid
ones cannot (e.g., a month is always related to its year). The relationships
shown in Fig. 5.26 are rigid, as indicated by the solid arrowhead.
Fig. 5.27 Browsing the hierarchy in the Time dimension
Figure 5.27 shows some members of the Calendar hierarchy. As can be seen,
the named calculations FullSemester , FullQuarter ,and FullMonth are displayed
when browsing the hierarchy.
The definition of the fact dimension Order follows similar steps than for
the other dimensions, except that the source table for the dimension is the
fact table. The key of the dimension will be composed of the combination
of the order number and the line number. Therefore, the named calculation
OrderLineDesc will be used in the NameColumn property when browsing the
dimension. Also, we must indicate Analysis Services that this is a degenerate
dimension when defining the cube. We will explain this in Sect. 5.9.5 .
Finally, in many-to-many dimensions , like in the case of City and
Employee , we also need to indicate that the bridge table Territories is actually
defined as a fact table, so Analysis Services can take care of the double-
counting problem. This is also done when defining the cube.
5.9.4 Hierarchies
What we have generically called hierarchies in Chap. 4 and in the present
one in Analysis Services are denoted as user-defined or multilevel hierarchies.
Multilevel hierarchies are defined by means of dimension attributes, and
these attributes may be stored in a single table or in several tables of a