Therefore, a key issue is the data warehouse
efficiency. Well developed solutions to this problem are based on materialized
views and query rewriting as well as on advanced index structures.
A challenging issue within the first solution concerns the selection of such a set of
materialized views that: (1) will be used for optimizing the greatest possible number
of the most expensive queries and (2) whose maintenance will not be costly. Several
research works have addressed this problem and they have proposed multiple algorithms
for selecting optimal sets of materialized views for a given query workload
(e.g., de Sousa & Sampaio, 1999; Gupta, 1997; Theodoratos & Xu, 2004).
A specific characteristic of OLAP queries that typically join fact tables with multiple
dimension tables as well as a specific distribution of values in fact tables requires
different indexing schemes. Three kinds of indexes have been developed in order
to optimize OLAP queries, namely, join indexes, bitmap indexes, and bitmap join
indexes (e.g., Aouiche, Darmont, & Boussa??d, 2005; O??™Neil & Graefe, 1995; Valduriez,
1987; Wu, Otoo, & Shoshani, 2004). The efficiency of executing OLAP queries
x
can also be increased by parallel processing and data partitioning techniques (e.g.,
Furtado, 2004; Rao, Zhang, Magiddo, & Lohman, 2002; St?¶hr & Rahm, 2001).
The ETL and refreshing processes may insert erroneous or inconsistent data into
a DW.
Pages:
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27