Common Sub-Expression Identification Strategy for MQO

Abstract

In query optimization, a query can be executed with different strategies, known as execution plan. The query optimizer will determine the best execution plan for a single query. However, when there are more than one query to be executed together, the locally optimal strategies for single queries may not be the best choice for obtaining a globally optimal execution cost. This will require for a multiple query optimizer that is able to select an alternative plan for each query in order to obtain an optimal global execution plan for the multiple queries. An optimal global plan can be obtained in two ways, by using an admissible heuristic with the search algorithms or by decreasing the number of search space through reducing the number of alternative plans generated. In generating alternative plans, the number and quality of the alternative plans produced are the major factors that will determine the performance of multiple-query optimization. In this paper, we propose the sharing opportunities for identifying the common sub-expressions.