Data Warehouse Concept Plaza W.H.INMON defines the data warehouse in the "Building Data Warehouse" is: Data warehouse is the topic, integrated, unrecognizable (stability), converged data collection over time. Therefore, it is generally considered that the data warehouse has four basic features: data warehouse data is topic; data warehouse data is integrated; data warehouse data is unrecognizable; data warehouse data is changing over time.
The theme is the analysis object in the field of analysis, and the topic's extraction should be determined according to the requirements. For example, what is the topic: the purchased subsystem, inventory system, sales subsystem in the MIS system, then the main items, suppliers, suppliers, suppliers, customers, etc. to be analyzed in DSS Therefore, the data warehouse corresponds to the topic of goods, customers, suppliers, etc.
Integration refers to the data in the warehouse is extracted from the originally dispersed database. There are many jobs to do during the data integration. For example, to remove noise data, it is clearly unreasonable data; there is also a place in all contradictions in the unified source of data, such as the unity of the field name, uniform unit, etc., is integrated before the data enters the data warehouse. For example, the original daily data is synthesized by month.
Unrecognizable means not updating data. Because the data of the data warehouse is mainly used by decision analysis, the data operations involved are mainly data queries. However, the no update in the inner is not the operation of Update, not the addition and deletion of data.
Data is constantly changing over time. Data warehouse changes with time to increase new data content, and continue to delete old content. The data warehouse contains a lot of time-related data, and the data should be re-synthesized over time. For example, this year's data week is integrated next year, it is necessary to integrate data quarterly.