View Relevance Driven Materialized View Selection in Data Warehousing Environment

A data warehouse uses multiple materialized views to efficiently process a given set of queries. These views are accessed by read-only queries and need to be maintained after updates to base tables. Due to the space constraint and maintenance cost constraint, the materialization of all views is not possible. Therefore, a subset of views needs to be selected to be materialized. The problem is NP-hard, therefore, exhaustive search is infeasible. In this paper, we design a View Relevance Driven Selection (VRDS) algorithm based on view relevance to select views. We take into consideration the query processing cost and the view maintenance cost. Our experimental results show that our heuristic aims to minimize the total processing cost, which is the sum of query processing cost and view maintenance cost. Finally, we compare our results against a popular greedy algorithm.