In these frameworks, getting rid of heterogeneity becomes a problem that appears
in the design and preprocessing stage of OLAP data. We explain a different strategy
to approach the problem, which is the model heterogeneity at the schema level of
the dimensions by enriching it with a class of integrity constraints called dimension
constraints introduced in previous work (Hurtado & Mendelzon, 2002; Hurtado
et al., 2005). Dimension constraints are Boolean expressions over categories and
paths of them, which allow stating restrictions on the structure of the dimension
elements. We explain two major applications of dimension constrains. We show
the role dimension constraints play in supporting transformations of heterogeneous
dimensions to the structural adaptations. We also explain how standard OLAP aggregate
navigation can be extended to heterogeneous schemas by performing inference
over dimension constraints.
Finally, we present the conclusions and outline open problems.
Structural. Heterogeneity
We will start this section by studying dimensions, facts, cube views, and data cubes,
which are the main notions in the multidimensional data model. We also define
structural heterogeneity. Then we will explain the notion of aggregate navigation
and the problems caused by heterogeneity.
Dimension.Modeling
In this section we introduce a model of OLAP dimensions presented in previous work
(Hurtado et al.
Pages:
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99