We explain the central concepts of the graph model proposed in previous
work (Hurtado et al., 2005), and study structural heterogeneity in this framework.
We explain why structural heterogeneity weakens the role of the hierarchy schemas
in supporting aggregate navigation, and in helping users to formulate correct OLAP
operations. We also explain the implication of heterogeneity to star and snowflake
models.
Figure 4. A heterogeneous version of a product dimension; (a) hierarchy schema;
(b) hierarchy domain
Figure 5. Different structures mixed in a heterogeneous dimension of Figure 4
(a) (b)
4 Hurtado & Gutierrez
Copyright ?© 2007, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of
Idea Group Inc. is prohibited.
We study different solutions proposed to keep the ability of OLAP systems to perform
aggregate navigation when the dimensions are structurally heterogeneous. We
overview the modeling of heterogeneity by the unbalanced dimensions of Jagadish
et al. (1999) and the solution proposed by Kimball (1996). We also explain how the
irregularity of heterogeneous dimensions can be fixed with null elements, which is
the approach of Pedersen, Jensen, and Dyreson (1999).
The solutions have in common that heterogeneous data are transformed into homogeneous
data; thus any commercial OLAP system may be used to handle the data.
Pages:
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98