基于互关联后继树的XML索引技术.pdfVIP

计算机研究与发展 ISSN1000—1239／cN11—1777／TP of Researchand JoumaJComputer Development 基于互关联后继树的XML索引技术 雷向欣 胡运发 杨智应 刘 勇 张 凯 (复旦大学计算机与信息技术系 上海200433) (1eixx@163．net) XML BasedonIRST IndexingTechnology Lei Kai Xiangxin，HuYunfa，YangZhiying，LiuYong，andZhang (D已加rf，”P行￡6，GD优户“￡i扎g丑珂dj以，br研口￡ioH丁童c^行。厶)g)，，F“(k，z【砌i说，百i砂，S矗n竹．g^口i200433) AbstractAnew schemefor r00tedtreesbasedonleaforderinterval scheme numbering labeling numbering Thenodesinarootedtreecanbeencodedinoncetraversalofthe andthe (LOINS)is tree， proposed． nodescanbedeterminedjnconstanttimebasedon ancestor—descendantshipamong IsBaRTI—I，anovelindexforrootedtreestructuredata whichtakesthe of model，isproposed advantages as and versionof IRST，suchindexingcompressibility．IsBaRTI—II，the IsBaRTI—I，is spaceoptimization alsointroduced． I indexesthe nodesandtheLOINSnumberof IsBaRTI—I，I ancestor—descendantshipamong the the and of inthe tree．Inthis node name(1abel)ofnode thecountits rooted by appearance way， structureand schemebecomesaunit and result indexing numbering unity．Theoryanalysisexperiment illustratesthat I needsmorelittletimeand tobe thenodeseriesand IsBaRTI—I，I capacitybuilt； path XPath canbeobtainedmore thanthe XML1ndexes matchingexpressions quickly pre