On view processing for a native XML DBMS

Abstract

XML is becoming the standard data exchange format. View transformation of XML data is important and frequent operation in XML data integration and publishing. In schema-based view transformation, users only need to define view schema over source to obtain view result. This declarative approach alleviates user from writing complex scripts to perform view transformation. Current available XML schema formats are too restrictive to express views with different semantic constraints. In this thesis, we introduce a semantically expressive XML data model: Object-Relationship-Attribute model for Semi-Structured data (ORA-SS), which allows users to define XML views with rich semantic meanings. Combined with ORA-SS, we use a native XML DBMS XBase which employs a new Object Based Clustering (OBC) XML storage scheme to perform accurate and efficient XML view transformation. Our experiments result shows our system is much faster than current XML query engines.