Improving Expressive Power in Modeling Data Warehouse and OLAP Applications - Evolving Application Domains of Data Warehousing and Mining - page 31

in different product categories. To represent this
kind of hierarchy in the multidimensional schema,
Malinowski and Zimányi (2008) define n on-strict
hierarchies as simple hierarchies that have at least
one many-to-many parent-child relationships at
the schema level; a hierarchy is called strict if all
relationships have many-to-one cardinalities.
Figure 3 includes the Assignation hierarchy
where an employee can belong to several de-
partments. Since, at the instance level, a child
member may have more than one parent member,
the members form an acyclic graph (Malinowski
& Zimányi, 2008). Non-strict hierarchies induce
the problem of double-counting measures when
a roll-up operation reaches a many-to-many

relationship, e.g., if an employee belongs to two
departments, his/her sales will be aggregated to
both these departments, giving incorrect results.
To avoid this problem, one of the solutions 7 is to
indicate, using the distributing factor symbol, that
measures should be distributed between several
parent members.
The mapping to the relational model provides
the same solution as presented in Figure 2: the
DimEmployee , DimDepartment , and EmplDep-
Bridge tables represent the Employee , Department
levels and many-to-many cardinalities with the
distributing factor attribute, respectively. How-
ever, it should be noticed that the logical-level
representation with three tables does not allow

Figure 11. A non-strict hierarchy in SSAS in a) a Dimension Usage and b) defining a many-to-many
relationship