Efficient support for XML queries on data base and data stream management systems.

摘要

The eXtensible Markup Language (XML) is rapidly becoming the de facto standard for retrieving and exchanging web information, and is having a significant impact on the evolution of existing database management systems. Commercial SQL-compliant data base systems are moving aggressively to support the integrated management of XML documents and XML-published relational data. This is achieved by supporting XML query languages, such as XQuery, along with extensions of relational languages such as SQL/XML.;In this dissertation, we focus our work on three areas that present research issues of particular significance and interest. The first area involves managing and querying the temporal history of databases using XML, since its data model and query languages are more supportive of temporal information than relational databases and SQL. We then discuss the problem of supporting this XML historical view efficiently by mapping back to relational databases, since these achieve better scalability performance. We finally introduce an algorithm which can process temporal coalescing efficiently, within the current SQL framework.;The second and fast growing opportunity area is represented by data streams, whereby query languages need to be extended to support continuous query applications on incoming relational streams and XML streams. While current approaches design separate data stream management systems for relational and XML data, the focus of my research has been on providing a unified support for both kinds of data streams. Toward this goal, XML SAX event streams and relational streams need to be integrated, and SQL and XQuery must be extended to overcome their limitations with certain data stream queries. We also overview the state-of-the-art technologies based on Finite State Automata (FSA) model to process multiple queries, and use UDAs to simulate such FSAs with comparable performance. These techniques allow us to build a system which integrated the management of relational and XML streams.;The last but not the least research problem in my thesis is OLAP applications on XML data. With XML gaining importance as the standard for representing business data, XQuery must support the types of queries that are common in business analytics. One such class of queries is OLAP-style aggregation queries. We review several recent research proposals which introduce new constructs into XQuery, pinpoint their disadvantages, and bring forward our function import mechanism. Basically, we allow XQuery UDFs to import functions written in SQL:2003 and UDAs. All complex moving window queries and grouping set queries can thus be specified in SQL easily using current SQL standards, and optimize using the mature query optimization technology of relational DBMSs. We show our approach provides a clear language expressive power and efficient query performance.