Abstract
XML becomes the standard for the representation of structured and semi-structured data on the Web. Relational and object-relational database systems are a well understood technique for managing and querying such large sets of structured data. Using an object-relational data model and an XML datatype, we show how a relevant subset of XML documents and their implied structure can be mapped onto ...
Abstract
XML becomes the standard for the representation of structured and semi-structured data on the Web. Relational and object-relational database systems are a well understood technique for managing and querying such large sets of structured data. Using an object-relational data model and an XML datatype, we show how a relevant subset of XML documents and their implied structure can be mapped onto database structures. Besides straight-forward mappings, there are some XML structures that cannot be easily mapped onto database structures. These structures would sometimes result in large database schemas and sparsely populated databases. As a consequence, such XML document fragments should be mapped onto database attributes of type XML and kept as is. The XML datatype implementation should support evaluating path expressions and fulltext operations. We present an algorithm that finds a type of optimal mapping based on the XML Document Type Definition (DTD) and statistics. The statistics are derived from sample XML document sets and some knowledge about queries on XML document collections.
Altmetric