Design Non-recursive and Redundant-Free XML Conceptual Schema with Hypergraph (Extended Abstract)

摘要

Data Type Definition(DTD) and XML Schema Definition(XSD) are the logical schema of an XML model, but there is no standard format for the conceptual schema of an XML model. Conceptual modeling is a very important first step for constructing a database application. A conceptual model describes a system that is being built. Abstract ideas are made concrete as the ideas are represented in a formal notation. A formal conceptual model has a number of advantages. First, it helps designers understand and document the application under construction. Second, it facilitates development of algorithms that derive the underlying database schemas. In this paper, a real world of interest is described in a conceptual-model Hypergraph, which is a generic conceptual model. It is a Hypergraph because its hyperedges, or simply edges, are not necessarily binary. Its vertices represent sets of objects and its edges represent relationships among the vertices. Edges in a Hypergraph can be directed or undirected, depending on whether the underlying relationships are functional or non-functional. As opposed to relational databases, in this paper we are interested in constructing XML database applications with 'good' properties. Two properties are particularly outstanding. First, the database should not have redundant data because redundant data lead to multiple-update problem once a single copy is modified. Second, since joins are expensive, the number of generated scheme trees, which are a generic hierarchical storage structure, should be as few as possible in order to reduce the number of joins required to answer a query. Users can draw a Hypergraph as XML conceptual schema with data relationships among elements as a result of specified functional dependency and multivalued dependency.