Indexing XML Documents
Indexování XML dokumentů
Authors
Supervisors
Reviewers
Editors
Other contributors
Journal Title
Journal ISSN
Volume Title
Publisher
České vysoké učení technické v Praze
Czech Technical University in Prague
Czech Technical University in Prague
Date
Abstract
Výzkum v oblasti indexování řetězců má již mnoho prezentovaných výsledků, což však neplatí pro ostatní datové struktury, jakými jsou například stromy.
Tato práce obsahuje v prvé řadě shrnutí metod pro indexování řetězců a stromů. Dále se podrobně zabývá rešerší existujících řešení indexování XML dokumentů.
Představena je zde nová jednoduchá metoda využívající deterministický konečný automat, jež umožňuje efektivně zpracovat XPath dotazy skládající se z libovolné kombinace child (/) a descendant-or-self (//) os, sloužících k navigaci v XML dokumentu. Spolu s touto metodou byly dále navrženy dva další konečné automaty na podporu jednodušších dotazů obsahujících vždy pouze jednu z uvedených os.
Ke konstrukci indexu pro daný XML dokument D s n elementy je využit odpovídající XML stromový model T. Zpracování dotazu Q o m elementech proběhne v čase O(m) nezávislém na n. Výsledkem dotazu je poté množina elementů splňujících dané požadavky.
Ačkoli automat podporující všechny dotazy s // osou indexuje až O(2^n) různých dotazů, počet stavů vlastního deterministického automatu je O(h^k), kde h je výška XML stromového modelu T a k je počet listů T. Pro běžné XML dokumenty lze navíc tuto mez triviálně snížit až na O(h.2^k).
The theory of text indexing is very well-researched, which does not hold for theories of indexing other data structures, such as trees for example. In this thesis we review existing techniques for indexing texts and trees and study state-of-the-art methods for indexing XML documents. We show that automata can be used effectively for the purpose of indexing XML documents. A new and simple method for indexing XML documents using deterministic finite automaton is introduced. The presented method supports a significant fragment of XPath queries which may use any combination of child (/) and descendant-or-self (//) axes. We also propose another two indexing techniques based on finite automata, aimed to assist in evaluating paths queries with either / or // axis only. Given a subject XML document D and its corresponding XML tree model T with n nodes, the tree is preprocessed and the index is constructed. The searching phase uses the index, reads an input query Q of size m and computes the list of positions of all occurrences of target nodes of Q in T. All the proposed automata performed the searching in time O(m) and do not depend on n. Although the automaton that supports all linear XPath queries where just // axis is used evaluates O(2^n) distinct queries, number of states of the deterministic automaton is O(h^k), where h is the height of T and k is the number of its leaf nodes. Moreover, we discuss that in case of indexing a common XML document the number of state in the deterministic finite automaton is at most O(h.2^k).
The theory of text indexing is very well-researched, which does not hold for theories of indexing other data structures, such as trees for example. In this thesis we review existing techniques for indexing texts and trees and study state-of-the-art methods for indexing XML documents. We show that automata can be used effectively for the purpose of indexing XML documents. A new and simple method for indexing XML documents using deterministic finite automaton is introduced. The presented method supports a significant fragment of XPath queries which may use any combination of child (/) and descendant-or-self (//) axes. We also propose another two indexing techniques based on finite automata, aimed to assist in evaluating paths queries with either / or // axis only. Given a subject XML document D and its corresponding XML tree model T with n nodes, the tree is preprocessed and the index is constructed. The searching phase uses the index, reads an input query Q of size m and computes the list of positions of all occurrences of target nodes of Q in T. All the proposed automata performed the searching in time O(m) and do not depend on n. Although the automaton that supports all linear XPath queries where just // axis is used evaluates O(2^n) distinct queries, number of states of the deterministic automaton is O(h^k), where h is the height of T and k is the number of its leaf nodes. Moreover, we discuss that in case of indexing a common XML document the number of state in the deterministic finite automaton is at most O(h.2^k).