Design And Implementation Of The Guangdong Tobacco ETL Of Decision Support Systems

In this paper, the background of Guangdong Zhongyan Decision Support System. The main source of data one project under the National Tobacco Board data, MES command systems, ERP systems and unstructured data. The main purpose is to establish a structured data for decision support systems environment through the historical data on the data source ETL process.At present, the enterprise information system was integrated multiple interfaces,but these systems are still relatively independent, different standards of design and development, the lack of a uniform horizontal data exchange mechanism, the formation of a " vertical " like information isolated island. Even the same indicator in different system performance for different data,"the number of departments, the value of diversity" brought difficulties to the development and utilization of information resources, give full play to the value of information resources more affected, leading to the existing data is difficult to directly applied to the auxiliary leadership decision-making. The systems that support leadership decision-making information systems to show the decision-making concern. This system will plurality of systems of the data extraction processing integration, is conducive to the data collector when the material of the various departments reporting. The system every day to update the latest data of the day and for processing, to solve the problem of data lag. In the first phase of the project, the main part of the completion of the decision-making center, by showing the focus of the decision-making level, to achieve the purpose of decision support.In this paper, through the ETL of the source data extraction, transformation, cleaning, loading a storage of high quality data of ODS (operational data store), through the ODS ETL establish data warehouse. Entire ODS divided into three layers: the interface layer, an intermediate layer aggregate layer. The interface layer of the data are consistent with the external system, The first is easy to extract data from the source system data to the ODS table to table.The second is reduce data processing in the data extraction process, convenient source table data verification, an error occurs in the data to reduce the complexity of the problem Find. The interface layer of the data to maintain real-time synchronization with the external system should be possible to improve the real-time nature of the data for real-time requirements of data fields. The middle layer is the core layer of the ODS store data, the data in principle, in the middle layer is the unified encoding format data, as the enterprise data standards guidance the peripheral system gradually unified data format. The data stored in the middle layer is the source of data warehouse data. Summary layer according to the processing needs of the subject domain, the formation of aggregated data, and loaded into the data warehouse.The Guangdong Zhongyan Decision Support System has been successfully on the line, about60%of the accounts the work of ETL development effort of decision support system is currently running in good condition.