Consider the following query that computes the maximum amount
over all the total amounts sold for each brand:
SELECT Brand, MAX(Amount)
FROM CV[Brand].
The result would be incorrect if the total sales at some null element surpass the sales
of each brand. This can happen in the dimension at hand if the sales of musical
products surpass the sales for each electrical product. Lehner et al. (1998) report
additional anomalies in actual OLAP scenarios due to contradictory queries that
arise when handling heterogeneity with nulls.
Structural.Adaptations
Another adaptation of standard dimension models to overcome the problems posed
by structural heterogeneity is proposed by Jagadish et al. (1999). The main property
of this model is to allow hierarchy paths starting from different bottom categories.
In this setting the hierarchy paths look unbalanced, from where comes the motivation
for calling so dimensions having this property. This model is subsumed by
Figure 8. A version of the dimension of Figure 4 with null elements: (a) hierarchy
schema; (b) hierarchy domain
(a) (b)
Handling Structural Heterogeneity in OLAP 4
Copyright ?© 2007, Idea Group Inc. Copying or distributing in print or electronic forms without written permission
of Idea Group Inc. is prohibited.
the graph model described in this chapter. The term unbalanced dimension is also
used to refer to dimensions which handle child/parent relationships that cannot be
modeled with fixed length paths in the dimension hierarchy (a good example is the
relation boss/employee in a company).
Pages:
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118