SEARCH
0-9 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Prev | Current Page 102 | Next

Robert Wrembel and Christian Koncilia

"Data Warehouses and Olap: Concepts, Architectures and Solutions"

Due to these problems, some
researchers have stated that the star schema does not allow structural heterogeneity
(Jagadish et al., 1999).
Table 1.
Product Brand Category Department
p1 b1 c1 d1
p2 b1 c2 d1
p3 b2 c1 d1
p4 b2 c2 d1
p5 b3 c2 d1
44 Hurtado & Gutierrez
Copyright ?© 2007, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of
Idea Group Inc. is prohibited.
Implications to the Representation of the Hierarchy Domain
In the presence of heterogeneity there is no precise correspondence between dimensions
and star dimensions. The hierarchy domain (i.e., the hierarchical arrangement
of elements) of a heterogeneous dimension may not be correctly captured by the star
dimension. As an example, if we represent the heterogeneous dimension of Figure 4 as
a star dimension, we cannot recover the original child/parent relation back. Indeed, if
we try to recover the rollup relation ?“[Brand,Category], we obtain the relation with a
single pair [Brand: b3,Category:c2]. The original rollup relation is show in Table 2.
Among other problems, this implies that the standard semantics of drilldown and
rollup operations differ for both models. Notice also that in this case, although the
dimension is heterogeneous, its star representation does not have null values. This
situation illustrates that, as explained in the introductory section, heterogeneity is not
only caused by the nonapplicability of attributes (e.


Pages:
90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114