As a consequence, user analyses will produce confusing or wrong results.
That, in turn, may result in disastrous business decisions made by decision makers.
For these reasons, research focused also on measuring and assuring data quality
(e.g., Jarke, Jeusfeld, Quix, & Vassiliadis, 1998; Vassiliadis, Bouzeghoub, & Quix,
1999).
For a long period of time, the existing DW technologies have tacitly assumed that
a DW is time invariant; that is, its schema and the structure of dimensions do not
change during a DW lifetime. In practice, however, a DW structure changes as the
result of the evolution of EDSs (Rundensteiner, Koeller, & Zhang, 2000), changes of
the real world represented by a DW, new user requirements, as well as the creation
of simulation environments, to list the most common cases. Several approaches to
handling changes in DWs have been developed. They are categorized as supporting
DW evolution (e.g., Blaschka, Sapia, & H?¶fling, 1999), temporal extensions (e.g.,
Hurtado, Mendelson, & Vaisman, 1999; Eder, Koncilia, & Morzy, 2002), simulation
(e.g., Balmin, Papadimitriou, & Papakonstanitnou, 2000) as well as versioning (e.g.,
Body, Miquel, B?©dard, & Tchounikine, 2002; Golfarelli, Lechtenb?¶rger, Rizzi, &
Vossen, 2004; Morzy & Wrembel, 2004).
In order to work properly and efficiently, all the above mentioned issues and techniques
need to use metadata.
Pages:
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28