Reverse engineering XML database into data type definition graph for data model verification

Abstract

Although XML became an industry standard for the data exchange on the World Wide Web, the most operational business data is stored in relational database systems That situation is likely to remain the same in the foreseeable future [10]. So we expect a lot of relational data to be converted or published as XML data. Up to now, a lot of researches have been done on translating the relation data into XML document and vice versa. But most works are only focused on the structure and neglected the semantic constrains. So some semantic constraints are lost after the conversion. However, the data semantics are important for database design. The main purpose of my final year project is to retain the relational database semantics in DTD and XML files so that users such as programmers or customers can benefit the data format interchange on the Internet. Because of the structure of XML and DTD different from the relational one, they cannot contain all relational database semantics. To solve the problem, I added id and idref relationships and used the existing occurrence operators of DTD file to maintain those semantics in XML and DTD files. The output Data Type Definition graph could be acted as an XML conceptual schema, which helped the user to visualize DTD structure for better understanding. User could also get the semantic information from the semantic list for analysis. Furthermore, it validated the value of idref attribute of elements by browsing the XML document. The validation ensured the integrity of an XML document. The invalid idref values were shown in error log.