Based On The Xml Heterogeneous Database Integration And Realization

For a long time ,there exist a variety of databases which are heterogeneous in that they store different types of data,represent data differently,use different software to manage the data and run on different computer hardware. With the farther development of business collaboration and government-affair electronical publishing,the information systems often requires access to data from multiple such databases.Therefore,a new system architecture is required to make the information sharing from many heterogeneous databases,in a transparent way,while preserving their autonomy.XML has a strong capacity of describing structural and semantic structural information;Data integration course has a better expansibility with the use of XML. XML has emerged as the dominant standard data exchange format for Internet-based business applications,the extensibility of XML allows the creation of generic models that integrate data from different sources.The key to realize integration of heterogeneous databases on XML is to establish a mapping approach.At prensent,There are two mapping approach on XML document. The one is based on templet ,the other is based on model.The two approaches exist some localization,the first one only supports mapping from relational databases to XML,the second has some restrict to XML document.This thesis begins with a compare on traditional approaches of heterogeneous databases integration. We analyzed the present algorithms and models on transformation of XML document and other format data,presented an approach based on DTD Element-tree.in the approach,the Element-tree is used the mapping between XML document and columns in the databases provided by the user quickly. The XDWrapper was a transformation system on the XML document and the relational database data,which adopted the Element-tree approach.XDWrpapper provides XML global views over heterogeneous relational databases by publishing relational data as XML,and allows outer user to access these views directly or query them using XML query language. The system has a simple mapping rule and may shield the different data sources.