Research On Normalization For XML Database Under Incomplete Information Circumstances

XML has become information representation and data exchange standards on the internet, and it is widely used in areas such as the network services, e-commerce, electronic data interchange, scientific data representation, data modeling and analysis, intelligent and search engine. XML technology is also increasingly concerned, XML database management technology has been matured and improved continuously. There is a great deal of incomplete information in objective world, the database which can express and deal with incomplete information has more realistic application meaning and value. In order to describe real world, XML document should allow incomplete information, however, after incomplete information is introduced in XML document, data constraints of XML document need to be redefined, so the normalization theory of XML database under incomplete information circumstances can't directly apply the corresponding theory of XML database under complete information circumstances. The normalization theory is a core issue in the field of database, and likewise it is of great significance for preventing update abnormity, reducing storage space, ensuring data consistency and query optimization. However, there is no perfect theory of literatures about normalization for XML database under incomplete information circumstances. This dissertation investigates the normalization theory for XML database under incomplete information circumstances.Based on the path and XML Schema, the normalization for XML database under incomplete information circumstances is systematically investigated, main innovative contributions in this dissertation are summarized as follows:Firstly, in the study of inference rules for XML strong functional dependency, the paper define XML strong functional dependency in leaf nodes (called XSFD) and investigate the property of XSFD. The paper also presents inference rules of XSFD for reasoning about the implication of XSFD and shows that the inference rules are sound and complete for arbitrary XSFD.Secondly, in the study of XML strong inclusion dependency normal form, the paper defines XML strong inclusion dependency (called XSIND) and investigates the property of XSIND. Inference rules of XSIND are put forward and the soundness and completeness of XSIND are proved. Based on the definition of non-interaction between XSFD and XSIND, the determinant theorem of non-interaction between XSFD and acyclic XSIND is presented; the paper defines XSIND normal form, studys the corresponding determinant theorem and presents the arithmetic of the normalization.Thirdly, in the study of the normalization for XML document with XML strong multivalued dependencies, the paper defines XML strong multivalued dependencies (called XSMVD) whose left path and right path are a single path. Based on the hierarchical XSMVD, the condition of satisfying XSMVD normal form for the incomplete XML document tree is proposed; the theorem that ensures redundancy-free in the incomplete XML document tree and the arithmetic for normalizing an incomplete XML document tree are put forward.Fourthly, in the study of the normalization for XML Schema with XSFD, the paper defines XML Schema, incomplete XML document tree according with XML Schema, XSFD in leaf nodes and branch nodes, etc. In order to solve logical implication, inference rules for XSFD are put forward and the arithmetics of path set strong closure and membership problem are proposed. For removing data redundancy, the definition of XSFD normal form is formalized and the corresponding normalization arithmetic is presented. Finally, in the study of the normalization for XML Schema with XSMVD,based on XML Schema, the paper defines XSMVD whose left path and right path are path set and investigates its property for XSMVD. Inference rules for XSMVD are given, the soundness and completeness of inference rules are proved. In order to removing data redundancy, the paper defines weak keys and XSMVD weak normal form, analyzes the reasons for data redundancies aroused by XSMVD in XML Schema via the instance and presents transforming rules and normalization arithmetic.The research in this dissertation gains the canonical incomplete XML document and schema by the normalization for the incomplete XML documents and schema in the web world. Storage, integration, distribution and transmission for the canonical incomplete XML data avoid update anomalies, ensure data consistency on the internet, improve data quality and have important practical value in the storage efficiency, index design and quey optimization.