Optimising web queries using document type Definitions

Abstract

A document type definition (DTD) D defines the struc-ture of elements permitted in any web document valid with respect to D. F rom a given DTD D we show how to derive a number of simple structural constraints which are implied by D. Using a relational abstraction of web databases, w e consider a class of conjunctive queries which retriev e elements from web documents stored in a database D. F or simplicity, we assume that all docu- ments in D are v alid with respect to the same DTDD. The main contribution of the paper is the use of the con- strain ts deriv ed fromD to optimise conjunctive queries on D by removing redundant conjuncts. The relational abstraction allows us to sho w thatthe constraints de-rived from a DTD are equivalen t to tuple-generating and equality-generating dependencies which hold on D.Having done so, we can use the chase algorithm to show equivalence betw een a query and its reduced form.