The materialized view technology has been widely adopted in the data warehouse systems to improve query performance. Two key underlying techniques are materialized view incremental refresh and query rewrite. However, due to the potential complexity of the materialized view defining query and dependency on the materialized view change log, not all materialized views are incrementally refreshable or generally query rewritable. Many restrictions need to be applied and addressed which make the materialized view not easy to use. This paper presents Oracle's tuning mechanism/tool to facilitate the materialized view creation by automatically fixing and/or decomposing the defining query and addressing the change log requirements. The resulting materialized view implemented by a set of recommendation statements achieves the goals of incremental maintenance and broad query rewrite.
[1]
Surajit Chaudhuri,et al.
Database tuning advisor for microsoft SQL server 2005: demo
,
2005,
SIGMOD '05.
[2]
Sam Lightstone,et al.
DB2 Design Advisor: Integrated Automatic Physical Database Design
,
2004,
VLDB.
[3]
Surajit Chaudhuri,et al.
Database Tuning Advisor for Microsoft SQL Server 2005
,
2004,
VLDB.
[4]
Pete Smith,et al.
1 – Data Warehousing
,
2005
.
[5]
Jingren Zhou,et al.
Stacked indexed views in microsoft SQL server
,
2005,
SIGMOD '05.
[6]
Benoît Dageville,et al.
Automatic SQL Tuning in Oracle 10g
,
2004,
VLDB.