XML Indexing and Retrieval with a Hybrid Storage Model

Abstract.

XML DTD (Document Type Declaration) puts two distinctive entities (attribute and element content) together into one framework for representing different document features. The notion of attribute in the XML DTD is similar to the field representation in the database, whereas the element content corresponds to the full text. In this paper, we view these two entities as different, each of which requires a different model for storage and retrieval. Attributes are stored in a database system, whereas the element contents and their indices are saved in files. We present a technique that puts together those two in an efficient way and builds an XML retrieval system on top of that. Such a system can achieve a reasonable trade-off between performance and cost in indexing and retrieval.