| With the popularity of information technology, Management Information System is more and more important to enterprise, it can provide data management, process planning, financial statement and other support functions, thus help administrators make management decisions. However, with enterprise growth, new data continues to be produced, which forms a massive accumulation in system, causes slow response to system queries and financial statements. This question can affect daily business activities, or lead to bad decisions in a further way.Materialized views (MV) as a technology to improve query efficiency, has been researched by many scholars. On one hand, MV calculates aggregating and join operations in advance, can improve system flexibility and query efficiency. On the other hand, it takes up certain space to save results and certain resources to refresh data. The focus of our research is how to select an appropriate MV set with the lowest total cost and apply them in management system, so as to get rapid query respond.The thesis takes ferry ticket management system as project background, verifies the necessity of query optimization through module analysis, and puts forward a MV-based query optimization model, which considers the query performance, update cost and storage space constraints. First, starts from the user’s target queries, through pre-processing and aggregating-operation analysis, gets characteristics of queries (that is, converts complex queries to unified simple ones). Then, divides them into different pedigrees, and finds candidate materialized views based on Aggregating Process Graph. Next, defines the formula of cost model, gain benefit and loss benefit, puts forward Mid-location algorithm which can select an optimal MV set through add views or delete views. Finally, applies the selected MV set in project system. Experiments show that the model has a good optimization performance, it can improve query response speed greatly and has huge practical application value. |
PDF Full Text Request