Research And Implementation Of Query Optimizing Of Column Store In Data Warehouse Management System

As a major study field of information technology area, there are many research findings on data warehouse technologies. The early researches on data warehouse were based on relational databases, that is, data warehouse constructions were on basis of some relational database management systems. But with the expansion of applications and the growing demands on performances, data warehouse based on relational database can't meet the current requirements any more. Taking the data warehouse as a stand-alone system and implementing a Data Warehouse Management System (DWMS) bottom-up with the adoptive techniques for data warehouse become a study aspect of data warehousing area.Featuring the data warehousing applications, views to change the underlying storage of data warehouse has been proposed. The most famous one is column store architecture. This storage architecture takes one column of a table as a data unit and stores it in the external medium, which has been proven to enhance the performances of data warehouse and OLAP applications. The reason is that the data warehouse queries are more concerned about the partial attributes of an entity rather than the all information of the entity.This paper mainly studies three issues, the architecture and function modules of DWMS, the implementations of column store technology and the query optimization research of column store. Based on the third part, this paper accomplishes the design and implementation of the query processing system in DWMS.In the first part, this paper studies the general data warehouse architecture and describes the functions of each component in data warehouse. And based on them, this paper presents the architecture of DWMS.In the second part, this paper describes the current study status of column store technique and its implementation. It designs the underlying storage engine in DWMS with the column store technique.In the last part, this paper studies the query optimization technology of column store. It puts forward the innovative index structure, AG-Tree, based on column store and provides the comparison model on tuple materialization strategies. It also studies the optimization of query plan, and based on which, the paper presents the design and implementation of the query processor in DWMS.