Design, Optimization, and Evaluation of a Path Query System

I.P. Szanto · Research Repository (Delft University of Technology) · 2005

The model of the Universe of Discourse usually needs to be mapped onto the information technology environment chosen by the engineers. This mapping costs significant efforts, time and energy to construct, to use, and to maintain. All these efforts, time and energy can be saved, if the engineers choose an object data management system based on an object model powerful enough to be used to construct the model of the Universe of Discourse. The query language and the query subsystem of an object data management system based on a powerful object model must be different from other query languages and subsystems, because the underlying data model is different. The goal of this research was twofold: (1) to create a query language for such an object data management system, which encourages the formulation of semantically sound queries, supports the formulation of reusable queries, facilitates the usage of recursive queries through intuitive syntax and semantics, and (2) to verify the created query language by the design, optimization and evaluation of a query system whose primary interface is the created query language, and which enables flexible, online, user-defined consistency control of the entered data. This thesis gives an introduction to query systems, and to Perspective-DB that is an object data management system based on a powerful object model. Furthermore, this thesis gives insight into the development stages of the path query system: simple path expressions, active views, combination of path expressions, filtered path expressions, and recursive path expressions. It also describes the way the query subsystem is reused to extend the export and import facilities of the data management system. It features a functional description of the query system and of the parts of the data management system relevant to the query system, and it also handles implementation and maintenance details.

Read the paper · More papers on PaperTik