Research On The Query Optimization Technology Under Automatic Summary Table Which Is Based On Statistics Process Graph

With the further development of information technique in recent years, the world has entered into an information age. More and more enterprises take information management as a key project in their enterprise constructions. Project following up deepens working process and increases application system data. Therefore, database system and data warehouse system are becoming an indispensable part in people's daily work.In order to generating a useful report for management decision, data analysts often have to extract data, conduct query and statistics in the business database and data warehouse system. As the data capacity is always huge, join operation and aggregation operation will be engaged during theses queries. However, traditional method is a great inconvenience to the whole system for it usually cancels the access permission to database and data warehouse during querying. Taking a lot of time to work, it is unable to meet the requirement of quick response. Thus, using what kind of technique to improve the efficiency and achieve quick response becomes an urgent technical problem demanding immediate solution.This paper analyzes the concept of atomic viewpoint and granularity in the first place, distinguishing query from statistics with a clear definition. Then it points to find out all the dimensions in the database system and sort them according to certain rules, analyzing atomization of aggregation operation in details and programming statistics as a set containing filter conditions and aggregation operation. If the dimension hierarchy operation and aggregate operation in two statistics are equal, label them as the same pedigree. This clearly stratifies the statistical process. Finally, the candidate list will be targeted through statistics process graph and selection algorithm of Automatic Summary Table (AST). This system promotes efficiency through accessing these logically independent ASTs rather than basic tables. In addition, the paper presents some details of how to use this technique under development circumstance of.NET platform and SQL Server database. Throughout the online system operation, AST based on statistics process graph improves the system operation effectively, and solves the time problem that large capacity database system responses to report query.AST based on statistics process graph can optimize the query including a lot of multi-table joins and complicated aggregate function, which is a technique borrowing the idea from trading space for time of materialized views. Its feature is to distribute the complex query into each ordinary query. In order to avoid massive queries, most of the collection has been finished in daily settlement. It is not only a powerful extension of current query optimization theory, but also guidance to large data report query optimization technique.