Efficient materialized view selection as constrained evolutionary optimization in data warehouse

Abstract

Huge amount of data is generated every day in e-commerce transactions. To be able to digest and analyze discovered information, e-commerce applications need to employ efficient data warehousing and OLAP systems. In large business organizations, the reports generated by data warehouse, support the executive managers to take decisions effectively. A Data warehouse, aggregates data from various heterogeneous and distributed data sources (online transaction processing systems) to support multi-dimensional (MD) analysis and decision making. OLAP (Online Analytical Processing) queries fired by the database users, involve challenges such as the big size of the data warehouse, complex nature of the queries and high response time of queries. To solve these problems, a set of views is extracted from the base relation of the data warehouse and is stored as materialized views. These materialized views are selected, based on the userand#8223;s requirements (e.g., frequently asked queries, size of the cube etc.). Here a view is an aggregated fact, defined as a result of operating a relational operator on one or two base relation(s). The analytical queries are answered faster using these materialized views. newline