Patterns and Types for Querying XML Documents

Abstract

Among various proposals for primitives for deconstructing XML data two approaches seem to clearly stem from practise: path expressions, widely adopted by the database community, and regular expression patterns, mainly developed and studied in the programming language community. We think that the two approaches are complementary and should be both integrated in languages for XML, and we see in that an opportunity of collaboration between the two communities. With this aim, we give a presentation of regular expression patterns and the type systems they are tightly coupled with. Although this article advocates a construction promoted by the programming language community, we will try to stress some characteristics that the database community, we hope, may find interesting.XDuce [40, 39], the ¦ © ¥ ¡ © ! primitive of Duce [4], the ¢ " ¦ § © ¦ § primitive of Xtatic [31], while for other languages, for instance XSLT [22], the iterator is hard-coded in the semantics itself of the language.For what concerns deconstructing primitives, instead, the situation looks clearer since, among various proposals (see the related work section later on), two different and complementary solutions clearly stem from practice: path expressions (usually XPath paths [21], but also the "dot" navigations of Cω or Lorel [1], caterpillar expressions [12] and their "looping" extension [33]) and regular expression patterns [41].Path expressions are navigational primitives that pinpoint where to capture data subcomponents. XML path expressions (and those of Cω and Lorel in particular) closely resemble the homonimic primitives used by OQL [23] in the context of OODB query languages, with the difference that instead of sets of objects they return sets or sequences of XML elements: more precisely all elements that can be reached by following the paths at issue. These primitives are at the basis of standard languages such as XSLT and XQuery.More recently, a new kind of deconstruction primitive was proposed: regular expression patterns [41], which extends by regular expressions the pattern matching primitive as popularised by functional languages such as ML and Haskell. Regular expression patterns were first introduced in the XDuce programming language and are becoming more and more popular, since they are being adopted by such quite different languages as Duce [4] (a general purpose extension of the XDuce language) and its query language QL [5], Xtatic [31] (an extension of C#), Scala [54] (a general purpose Javalike object-oriented language that compiles to Java bytecode), XHaskell [45] as well as the extension of Haskell proposed by Broberg et al. [11].The two kinds of primitives are not antagonist, but rather orthogonal and complementary. Path expressions implement a "vertical" exploration of data as they capture elements that may be at different depths, while patterns perform a "horizontal" exploration of data since they are able to perform finer grained decomposition on sequences of elements. The two kinds of primitives are quite useful and the...

Optimizing XML querying using type-based document projection

“…In particular it seems worthwhile to consider more theoretically sound formalisms for tree queries such as, for instance, MSO formula or tree automata. The latter in particular would allow us to reuse our pruning algorithm for pattern-matching based languages (such as the CDuce language [1] and its pattern-based query language CQL [28,6,14]). It is also known that tree-automata (as well as MSO formula) have better closure properties than XPath expressions and support fine-grained set-theoretic operations (intersection, union, complement) that have been used with success to devise very precise type-systems for XML [22].…”

XML Typechecking

“…html X XAct: http://www.brics.dk/Xact/ Xduce: http://xduce.sourceforge.net XHaskell: http://taichi.ddns.comp.nus.edu.sg/ taichiwiki/XhaskellHomePage Xtatic: http://www.cis.upenn.edu/~bcpierce/ xtatic A gentle introduction to exact typechecking for both XML-to-XML transformations and XML Publishing can be found in [37]. A nontechnical presentation of Regular Expression Types and Patterns and their use in query languages can be found in the joint DPBL and XSym 2005 invited talk [4]. For a more complete presentation of Regular Expression Types and Patterns and the associated type-checking and subtyping algorithms we recommend the reader to refer to the seminal JFP article by Hosoya, Pierce, and Vouillon [22].…”

Semantic subtyping with an SMT solver

“…Frisch et al, (2008) extended semantic subtyping to function types and propositional types, with typetest, resulting in the language CDuce (Benzaken et al, 2003). (An excellent overview of the use of semantic subtyping in the context of querying XML documents was given by Castagna, 2005. ) In the end, the XQuery working group resorted to a more conventional pure named type system (Siméon & Wadler, 2003) with a simpler notion of subtyping based on ordinary regular expression inclusion (as opposed to XDuce's use of tree regular expressions).…”