Design And Implementation Of The Population Data Warehouse System For The Public Security Bureau

With the development of information technology, public security departments have accumulated a lot of historical data with less usage. The purpose of the data warehouse is to establish a system of data storage environment. It uses technologies of the online analytical processing (OLAP) to help users extracting useful information from large amounts of data and facilitates decision-makers make decisions on the application of this information. Thus the establishment of public security data warehouse can dramatically improve the above situation and raise the level of information.The main research directions of the data warehouse has the following aspects:the choice of subjects, ETL(extract, transform and load) operations, online analytical processing, materialized views'selection and maintenance, distributed data warehouse and data mining technologies based on data warehouse. Materialized views are views that store summary data in data warehouse for the purpose of improving query efficiency. This paper focuses on view selection and maintains, that is how to select part of summary results, store them as materialized views, and maintain them to keep data consistency on conditions of limited system resource.This paper introduces the data warehouse and OLAP of concepts, then describes the materialized view of the concept, advantages and disadvantages. After that this paper researches the materialized views'selection and maintenance technology. Selection technologies include static and dynamic view selection. For the purpose of reducing replacement frequency and the cost of I/O, it proposes an improved dynamic view adjustment algorithm which bases the greedy algorithm and dynamic selection algorithm using the mechanism of time stamp and validity. At the same time, this paper compares the current common view maintenance technologies and proposes a new method called TS-BT-DV which uses the timestamp, the buffer table and the double version of the control technology, as a result, it can guarantee users'querying and view maintenance in parallel, and avoid frequent updates of views. Experiments show that both the above two algorithms achieved good results. Finally, this paper introduces how to design police population data warehouse.