Efficient execution of data warehouse query using look ahead matching algorithm

Abstract

Data warehouse (DW) is always subjected to large and complex workloads of queries. Aggregate function computation and iceberg queries are important and common in many applications of data mining and data warehousing because people are usually interested in looking for unusual patterns by computing aggregate functions across many attributes. In spite of complexity of these queries decision makers want their request to be processed quickly. But these queries often require very long response time. So it is very important to process efficiently these expensive queries with aggregate functions in data warehousing environment. Presently available Iceberg query (IBQ) processing techniques faces the problem of empty Bitwise AND operation, and require more I/O access to execute query. None of the research provides the model for all aggregate functions. Proposed research applies look ahead matching strategy on Bitmap Index (BI) of query attributes. Before performing actual operation the analysis of logical operation is done if it satisfy threshold condition then only complete operation will be perform. In this way look ahead matching(LAM) strategy reduces the I/O access cost and solve the problem of empty bitwise AND operation. This research also proposes framework for other aggregate functions like MIN, MAX, SUM and COUNT.