Application Research Of Real View Technology In Agricultural Production Data Warehouse

Data warehouse is generated with the existence of large numbers of databases, and for further mining the data resources and meeting the needs of decision support. It is not the so-called "large database", it is a data set which is particularly designed and established to support business decisions. The data warehouse technologies are based on the business development needs of information systems and extended from the database system technology. Materialized view is one of the important technologies to improve the performance of data warehouse, by pre-storing the results of queries which are used frequently to speed up the query response time greatly.Materialized views include two important technologies, that is, materialized view selection and materialized view maintenance. Materialized view selection is a technology which is used to select suitable views to materialize among candidate views. When data sources update, maintenance algorithm is called in order to be consistent with the data sources. Therefore, main research contents of this paper center materialized view selection and materialized view maintenance.The research content in this paper mainly focuses on the following parts:1. Applied research in this paper is based on the project of "Agricultural quantity security intelligence analysis and early warning systems and demonstration of key technical support", so firstly we need to build an agricultural production data warehouse. The first step is the frame design of database and data warehouse, which includes the structure design of data tables and middle tables, and the partition of dimension and measure and so on. The second step is data cleaning, storing the data with unified organizational format, which is selected from the data sources, to the middle tables. The third step is the process of ETL, loading the data which is stored in the middle tables into the database using the self-programming ETL tool, then building the data warehouse using SSAS tool provided by SQL Server.2. By analyzing and summarizing the traditional materialized view selection algorithms, we can find out that the traditional algorithms are high complexity and slow query response. To overcome these drawbacks, this paper presents an improved genetic algorithm to solve the materialized view selection problem under query cost constraints. The algorithm dynamically changes the crossover probability and mutation probability in the process of genetic. In this way, it can not only maintain the population diversity, but also ensure the convergence of the genetic algorithm. So it effectively improves the optimization ability of genetic algorithm, thus avoiding the "evolutionary stagnation" problems. Meanwhile, the improved genetic algorithm increases the processing of invalid solution to avoid the "evolutionary stagnation" problems generated by invalid cycle. Thereby the efficiency of materialized view selection is greatly improved.3. By analyzing and summarizing the traditional materialized view maintenance algorithms, we can find out that the traditional algorithms have update abnormity problem. To overcome these drawbacks, this paper presents an improved materialized view maintenance algorithm which based on EC A algorithm. Experiments in this paper prove that the improved algorithm has solved update abnormity and guaranteed data consistency.4. The theoretical research results are applied to the agricultural production data warehouse, which proves the practical value of the improved materialized view selection algorithm and the materialized view maintenance algorithm.