DocumentCode :
1122671
Title :
View adaptation in the fragment-based approach
Author :
Bellahsene, Zohra
Author_Institution :
CNRS, Univ. Montpellier II, France
Volume :
16
Issue :
11
fYear :
2004
Firstpage :
1441
Lastpage :
1455
Abstract :
View adaptation relies on adapting a set of materialized views in response to schema changes of source relations and/or after view redefinition. Recently, several view selection methods that are based on materializing fragments of the view rather than the whole view have been proposed. We call this approach the fragment-based approach. This paper presents a view adaptation method in the fragment-based approach, which is aimed at exploiting the opportunities to share not only materialized data, but also computation between the different views. In order to do this, the views are modeled using the so-called multiview materialization graph, which represents the views as a bipartite directed acyclic graph whose nodes are operations and fragments of the views. Then, the adaptation is performed regarding all materialized views and not solely the old materialization of the view. However, the data independence is preserved for the views that are not affected by the change. On the contrary, in related work, the adaptation technique is based solely on the old materialization of the same view. We studied the impact of the fragmentation on the adaptation techniques and showed the advantages and drawbacks of this approach.
Keywords :
data warehouses; directed graphs; query processing; synchronisation; bipartite directed acyclic graph; data independence; data warehouse evolution; fragment-based approach; materialized view; multiview materialization graph; source relation; view adaptation; view redefinition; view selection method; view synchronization; Computer Society; Data warehouses; Warehousing; 65; Index Terms- Materialized views; data warehouse evolution; view adaptation; view redefinition; view synchronization.;
fLanguage :
English
Journal_Title :
Knowledge and Data Engineering, IEEE Transactions on
Publisher :
ieee
ISSN :
1041-4347
Type :
jour
DOI :
10.1109/TKDE.2004.79
Filename :
1339269
Link To Document :
بازگشت