Research and Comparison of SQL Optimization Techniques based on MapReduce

Abstract

MapReduce [1] is a software framework for distributed computing. It has been widely used in data processing and analysis. SQL-like queries are also playing important roles in data analysis. In this paper, we study different approaches to optimize SQL-like queries in MapReduce and evaluate them on benchmark dataset in a Hadoop cluster. The result shows that these optimizations including adjust join order, push down predicates and merge different MapReduce tasks can significantly speed up analytical queries.