Research And Realization Of Data Warehouse For Tobacco Marketing

Enterprise management is demanding determination to attain a higher level. The correctness and effectiveness of determination depend upon sustain of data. Data warehouse provides an integrate solution to data sustain. Generally speaking, Data warehouse's structure includes five levels : Data source, Monitor, integrator, Data warehouse and Client application. Course of establishing a data warehouse includes: 1.Concept model designing: Concept mode is induction from a real data warehouse. It indicates whether the designer understand his customers' demands or not. 2.Logic model designing: Logic mode is a bridge between concept model and physic model. 3. Physic model designing: Physic model is a model of realizing logic model in physic system. When a data warehouse has been established, pre-treated data must be loaded into it to make it operate. Course of data pre-treatment includes: 1.data sources selection. ;2. Extracting data from data sources ; 3. Washing and transiting data according standard format and notation of the data warehouse; 4.Loading data into the data warehouse. We can query and analyze data in data warehouse when data has been loaded. As data operating and exhibiting tool, online-analysis process OLAP brings great convenience to data warehouse query and analysis. OLAP has two models: MOLAP model and ROLAP model. The former uses multi-dimensional database to store data. While in the latter, data is stored as rows and columns that are expressed with relation forms, and exhibited to customers as business dimension. The main difference between MOLAP and ROLAP lies in: query capability, analysis capability, data storage capability, data loading capability, maintenance capability. Data scale is the main consideration for selection between MOLAP and ROLAP. Using SQL Server, this text realizes establishment of data warehouse , tobacco sales data query and analysis in the established data warehouse.