In this paper we address the problem of immediate translation of XPath+IR queries to relational database expressions and exert the benefits of using an intermediate algebra. Adding an intermediate algebra on the logical level of a database enables a level of abstraction from both query languages for IR in XML documents and the underlying relational storage. This paper proposes a region algebra that can be extended to support ranking operators in an elegant way while staying algebraic. Furthermore, region algebra operator properties provide a firm ground for query rewriting and optimization.
|Title of host publication||Proceedings of the Joint Workshops on XML, IR, and DB|
|Editors||Ricardo Baeza-Yates, Yoelle Maarek, Thomas Roelleke, Arjen P. de Vries|
|Place of Publication||Sheffield, UK|
|Number of pages||8|
|Publication status||Published - Jul 2004|
|Event||1st Workshop on the Integration of Information Retrieval and Databases, WIRD 2004 - Sheffield, United Kingdom|
Duration: 29 Jul 2004 → 29 Jul 2004
Conference number: 1
|Workshop||1st Workshop on the Integration of Information Retrieval and Databases, WIRD 2004|
|Period||29/07/04 → 29/07/04|
- DB-XMLIR: XML INFORMATION RETRIEVAL
Mihajlović, V., Hiemstra, D., Blok, H. E., & Apers, P. M. G. (2004). An XML-IR-DB-Sandwich: Is it Better with an Algebra in Between? In R. Baeza-Yates, Y. Maarek, T. Roelleke, & A. P. de Vries (Eds.), Proceedings of the Joint Workshops on XML, IR, and DB (pp. 39-46). Sheffield, UK.