Research On Equipment Data Modeling And Query Optimization Technology

With the development of the data project whose target is to construct data sets for demonstrational data of equipment,various data of equipment and related comprehensive data become rich and perfect,and the scale of data grows enormously.In the era of big data,data mining,decision support system and other applications of analysis technology have become mature,which make it possible to explorer the potential value of the demonstrational data of equipment.This paper focuses on the method of data representation when the scale of data becomes large and the structure of data organization becomes complicated.Materialized views have been used to improve the efficiency of query processing of DBMS,and cuckoo search base view selection algorithm was proposed.Firstly,according to the theme of management of full life of equipment,a multidimensional data model which based on the feature of demonstrational data of equipment and the practice of equipment management business was designed.The design process has been described in detail from conceptual model designing and logical model designing.The advantages of the multidimensional data set in data representation have been analyzed.Secondly,the technology of materialized view has been applied to improve the data query efficiency.The advantages and disadvantages of the classic model of view selection problem have been analyzed and compared on the basis of the multidimensional data set.Data cube model has been selected to solve view selection problem which was limited by the cost of maintenance and update.Improvement has been made upon classical linear cost model,the definitions of query cost,benefit after materialization and maintenance update cost have been proposed.Cuckoo search based materialized view selection algorithm has been proposed,and experiments on TPC-DS benchmark and self-build multidimensional data set have been performed.It has been verified that the new algorithm has good performance for solving view selection problem when the scale of data was gigantic and the organization of multidimensional data set was complicated.The new algorithm has universal applicability and the ability to solve practical problems.