The Solution Space Study Of Query Rewriting Based On Query Optimization

With the rapid development of the storage capability of the database, the query efficiency becomes more and more important. Query optimization is the key factor of RDBMS Query efficiency. Query optimizer process query optimization to SQL submitted according to system state and Stat, information and the intelligent degree of it determins the query efficiency of RDBMS.The synthesis of query cost mainly reflects the number that the disk is visited during the execution of database access. This paper puts forward a calculation approach of it, high query frequency leading directional move, based on Query relevance: creating a field to stat. query frequency on source table, creating fields to save message of query and other operations on stat table to record the main methods of access. When analysing aim table , we can recreate table and keep the high frequency queried records on the table head to increase the speed when the query module is first_row. The record of high query frequency should be kept on the table head on the condition of the high mobility of database because the last inserting records is more likely to be accessed. This paper gives a evidence of this viewpoint. Because the viewpoint of'high query frequency leading directional move'depends on the real_time stat. data that must be got by DBA's freehand analysis, we raise another method, intelligence query analyse agent, which gets CPU state, dynamically arranges analysing task and forms the query_plan for Query optimization.Join-query is the most important part of query optimization. A paper written by Jianping Peng and Bianqin Wang only mentions two states of'no branch join-query'and'star join-query'. But there is another way of"net join-query"in the practical usage. After the study of the above two approaches, this paper raises an advanced scheme: changing the'net join-query'into query connected graph, analysing the query connected graph and obtaining the compromise weight. According to'the minimum spanning trees'rules we can turn the'net query'into'tree query'in the least time, and bring forward a way to prove that this way have a least weight. This article analysed the elements that affect five kinds of caculation of join-query and pointed out the importance of creating suitable types of database.The article, taking a concern application system as an example, illustrated how a series of approaches such as create suit table, right sql etc .optimize enquiry at all levels.