Automated Database Restructuring: Using Views To Recast The Stored Data

In modern database systems, it is common to define views of stored data. Views can be materialized, that is precomputed and stored in the database. Views are typically materialized to improve some metric of database performance, such as query-evaluation costs or maintenance costs for the already materialized views. To optimize the chosen performance metric, it is critical to materialize the "right" views. The problem of finding views to materialize has been studied using the view-selection approach, where the views to materialize are given in advance. However, in general, it may be necessary to invent new view definitions. Automated database restructuring is the problem of automating the invention of new views. The database-restructuring methods presented in this book are applicable in many areas, including query optimization, data warehouse design, and information integration.