The Intelligent Optimization Of Complex Query Based On Cloud Database

In recent years,with the rapid development of network services,the exponential growth of data requires more storage space and computing resources,which brings some challenges to enterprises and users.Facing the rapid growth of data volume,the expansion of the traditional selfbuilt database needs to add a large number of hardware devices,and the deployment is very cumbersome and time-consuming.Compared with the traditional self-built database,cloud database has gradually become the choice of users and enterprises because of its high cost performance,high expansion and high recovery.However,tuning the cloud database creates difficulties for users.With the rapid development of artificial intelligence,many scholars have applied AI to the field of cloud database tuning with remarkable results.However,in many specific problems,such as the optimization of cloud database query,the research is not comprehensive enough,and there are still some problems that have not been solved.Therefore,based on DBMS automatic tuning technology and machine learning method,this paper carried out research on the optimization of cloud database related problems.This paper first studies the important operation Hash Join in the analysis task of cloud database,and implements an intelligent optimization scheme of Hash Join column row rotation.It is the first time to optimize the Hash Join operation in the query by using column row rotation and obtain huge benefits.The feasibility of column row-optimized hash Join is verified by theoretical derivation and data experiments,and the conditions for column row-optimized hash Join are explored.The intelligent decision model is also introduced in the scheme,and the idea of ensemble learning is used to automatically judge whether the column to row operation on Hash Join can get profits.The validity of the proposed method is verified by experiments.In an analytical cloud database,generating an optimal query plan requires obtaining statistical information(such as density distribution,row number,cardinality)on the combined dimensions of the queried data in each data column.The existing methods in the industry are often not ideal and may have negative effects on query plan optimization.The second part of this paper is based on this background,a database statistics collection and optimization system based on the historical query execution information and the intelligent decision model is designed and implemented.The statistics used in query optimization can be modified timely through the real statistics information after query execution,so as to provide accurate statistical information reference for subsequent similar queries without adding additional query load.Also introduced the intelligent decision module,the change of relevant statistical information and the impact on the query plan generated modeling study,detection and prediction of statistical information and the change of larger query execution performance differences,thus targeted before the execution of the query statistics collection,to improve the whole database query plan optimization,Effectively improve the efficiency of database query.