Selection of Frequent Queries for Constructing Materialized Views in Data Warehouse

A data warehouse contains archival data on which analytical queries are posed for decision making. Since most of these analytical queries are complex and are processed on a large data warehouse, the response time is high. This response time can be reduced by constructing materialized views containing information that has a high likelihood of answering future user queries. One way to identify such information is by using the previously posed queries on the data warehouse. It is generally observed that most of these previously posed queries access similar information and are posed more frequently as compared to a few rarely posed queries. The frequently posed queries, referred to as frequent queries, are appropriate for constructing materialized views. The approach presented in this paper attempts to select such frequent queries from all the previously posed queries on the data warehouse. The information in these selected frequent queries may have a high likelihood of answering future user queries and therefore, can be used to construct materialized views. As a result, most of the user queries will be answered by the materialized view, instead of the large data warehouse. This would reduce the query response time and thereby facilitate decision making.