A Classification Scheme for Semantic and Schematic Heterogeneities in XML Data Sources
Charnyote Pluempitiwiriyawej, Joachim Hammer · 2000
The integration of distributed, heterogeneous information sources has been the topic of intense investigations for many years. Most of the research in this context has focused on integrating well-structured data stored in relational or object-oriented databases. In recent years, the World Wide Web (Web) has become the single largest, heterogeneous source for almost any kind of online information. The Extensible Markup Language (XML) is well suited to represent and share this Web data which is mostly semistructured in nature. Using XML, data is stored in an XML document whose structure is defined by one or more Document Type Definitions (DTD). Despite XML’s support for the sharing and exchange of information, the integration of multiple, heterogeneous XML documents remains a challenge due to semantic and schematic discrepancies among the data. In this report, we propose a directed, labeled graph model to describe the internal structure of XML documents and their associated DTDs. Based on our model, we introduce a classification of the possible conflicts that can occur between data that is represented in XML. We believe that such a classification is a useful and important guideline for resolving conflicts that occur when sharing data represented in XML. 1