There are also atoms that restrict categories that may be indirectly reached by the
root category. As an example, the dimension of Figure 4 may be modeled with just
the following constraint:
??©Product,Brand??? ??? ??©Product,..,Category???.
Handling Structural Heterogeneity in OLAP
Copyright ?© 2007, Idea Group Inc. Copying or distributing in print or electronic forms without written permission
of Idea Group Inc. is prohibited.
With this constraint we discard structures that do not contain Brand or Category.
However, the product elements may reach both directly or indirectly the elements
at Category. In particular, the edge BrandCategory allows some elements to reach
Category indirectly passing through Brand.
In previous work (Hurtado et al., 2005), dimension constraints are compared with
other classes of constraints proposed to capture different forms of heterogeneity
for OLAP (Huseman et al., 2000; Lehner et al., 1998), relational (Goldstein, 1981),
and semistructured (Abiteboul & Vianu, 1999; Buneman, Fan & Weinstein, 1998)
data. Among them, the constraints of Husemann et al. (2000) are closely related to
dimension constraints in that they address a form of structural heterogeneity study in
this chapter. These constraints are subsumed by dimension constraints. They allow
expressing that two paths in the hierarchy schema that start from a single category
are mandatory or alternative
Extracting.
Pages:
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124