Jagadish et al. (1999) propose SQL(H),
a query language that includes the descendant/ancestor relationship as a built-in
predicate which allows for succinctly expressing many useful queries in an SQL
style. The rollup relation between two categories allows expressing relational queries
over the graph model of the dimension (e.g., using SQL). Let c1, c2 be two categories
of a fixed dimension such that c1 reaches c2 in the hierarchy schema (i.e., c1 ??°
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.
c2). The rollup relation from c1 to c2 in a dimension d, denoted ?“[c1, c2] is defined
as the relational table having attributes c1, c2, and containing the set of tuples [c1 :
e1 : c2 : e2] such that d(e1) = c1, d(e2) = c2, and e1 < e2. Notice that the rollup relation
is defined for pairs of categories that are (directly or indirectly) connected in the
hierarchy schema. Note also that the notions of strictness and homogeneity can be
defined in terms of the rollup relation. A dimension is strict if and only it has functional
(i.e., single valued) rollup relations. A dimension is homogeneous if and only
if each rollup relation ?“[c1, c2] is total over the elements in c1.
Facts and Data Cubes
A fact is a data entity that relates a list of elements, taken from a list of dimensions,
to measures of interest.
Pages:
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104