(2005).
Here, experiments show that Dimsat takes few seconds to test the implication of a
dimension constraint from a highly heterogeneous schema of 25 categories.
Supporting.Structural.Adaptations
The framework of dimension constraints and frozen dimensions can be used to automate
the construction of the structural adaptations explained in previous sections.
In this form we may use the semantics of the schema to transform heterogeneous
dimensions, a task that could be difficult to do manually when handling complex
dimensions. As an example, it is not easy to obtain the unbalanced dimension of
Figure 11 from the entangled product dimension of Figure 7.
Frozen dimensions correspond to the homogeneous dimensions in the decomposition
proposed by Kimball (1996), so the decomposition can be obtained by the algorithm
that computes frozen dimensions, explained before. In Hurtado and Gutierrez
(2004), we propose an algorithm that transforms heterogeneous dimension schemas
into the unbalanced homogeneous schemas of Jagadish et al. (1999), we called canonical
schemas. The algorithm first computes the set of frozen dimensions of the
original schema and then iteratively splits the categories that cause heterogeneity.
In each split operation the hierarchy schema is transformed so that the resulting
hierarchy schema has a single category for each different frozen dimension over
the category.
Pages:
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127