The Design And Implementation Of ETL Tool In Medical Case Resource Database Based On Metadata

It is urgent for the development of a hospital to construct a Data Warehouse that gives decision-support for ancillary clinical diagnosis and daily management.Oriented to requirement of the hospital, Medical Case Resource Database (MCRD) integrates mass of data and mines valuable information by mining techniques, and it is practical in applications such as clinical diagnosis, management, researching and teaching. The ETL(Extract Transform and Load, ETL)systems are responsible for the extraction of business data from several systems, processing the extracted data(such as cleaning, transforming etc.)and loading the processed data into the target MCRD.The paper aims to implement an ETL tool featuring MCRD by which data source is extracted and loaded to MCRD. First, based on the thought of metadata, its architecture is constructed by analyzing its requirements and the characteristics of data sources. Second, structure design of metadata is focused on and the processing flow in every module is sketched. Finally, ETL is achieved.And XML (Extensible Markup Language) is used to organize the store the metadata, temporary data and in cleaning and transforming.Metadata, as a technique, decreases largely the difficulty of the implement, increases the expansibility and flexibility of the system. The paper explores how to make the rules of metadata in files. Another exploration is that all the data is organized in the format of XML rather traditional database in the whole flow.The tool has already been run in the hospital, and it is testified in the usability, expansibility and stability.