A framework for designing materialized views in data warehousing environment

Abstract

Data warehouses are accessed by different queries with different frequencies. The portions of data accessed by a query can be treated as a view. When these views are related to each other and defined over overlapping portions of the base data, then it may be more efficient not to materialize all the views, but rather to materialize certain "shared views" from which the query results can be generated. We address some issues related to determining this set of shared views to be materialized in order to achieve the best combination of good performance and low maintenance, and provide an algorithm for achieving this goal.