Research On Multivalued Dependencies And Reference Rules For XML With Incomplete Information

In the real world, data stored in the computer system and the database system has n^n kinds of forms. To transmit data among systems on the Web costs developers a great deal of time. Converting data into eXtensible Mark-up Language (XML) format reduces the complexity of data transmission greatly, and makes the data read by various programs. XML is becoming a primary standard for data representation and transmission on the internet.Although it is very easy to represent data from different source for XML document, but its semantic information is relatively limited. Starting from XML Schema, this paper strengthens the ability of defining integrity restraint and semantic expression. The research is made step by step: at first, by analyzing the difference between Document Type Definition (DTD) and Schema, and referring to the relevant study on integrity of relational database and XML document based on DTD, this paper explored the theory of fundamental data dependencies in XML document based on Schema. To deal with the tree-like and semi-structured style of XML documents, definitions are given based on path expression. Next, referring to the theory of relational database with incomplete information, the concept of incomplete information is brought in XML document. The contributions of this paper are as following:First, XML functional dependencies definition and inference rules are given, which soundness and complements is proved. Algorithms of simplifying right path, the closure set of left path, and normalization cover are proposed.Second, the concept of key is given at the base of element type that takes part in closure operation, and its inference rules are given, which soundness is proved. Algorithms of transformation between absolute XML functional dependencies or key and relative one are proposed.Third, XML multivalued dependencies definition and inference rules are given, which soundness and complements is proved. Algorithms of extracting multivalued dependencies set from the schema, determining the biggest scope path, judging whether the certain multivalued dependencies set is redundant and normalization cover are proposed. Relationship between relative XML key and relative XML multivalued dependencies is studed. The solutions to resolve two typical redundant problems in XML documents are given.Forth, incomplete information is brought into XML document, the definition of strict and lax multivalued dependencies are given, and research on inference rules and cover set is made with semantic constraints respectively.The functional dependencies, key and multivalued dependencies can express not only the absolute integrity constraints in the whole document, but also the relative integrity constraints in part of the document.Realistic significance of this paper lies on:Fundamental data dependencies descript relationship between entities that element type and attribute type represent from the aspect of value; the base is given on which to study the problem of normal form, optimizing query, structure restraint and path restraint etc; comparing with DTD, XML document based on Schema is more fit for data transmission and conversion with relational database system; incomplete information in the XML document improves the ability of semantic expression.