{"title":"Efficient RDF Query Processing using Multidimensional Indexing","authors":"Akrivi Vlachou","doi":"10.1145/3139367.3139408","DOIUrl":null,"url":null,"abstract":"Efficient management of RDF data is important due to the increasing number of data sources available in RDF, which fosters interoperability and enables data integration and interlinking. An RDF store maintains data in the form of triples, subject (S), property (P), object (O). A common approach for storing RDF data is to follow a so-called \"one-triples-table\" approach, where all triples are stored in a single table. For efficient data access, multiple one-dimensional clustered indexes (for example B+trees) are built, practically replicating and storing the triples in different sort orders, thus enabling efficient random access based on any subset of SPO. An issue that is largely overlooked in related research is that since triples are 3-dimensional, a natural alternative is to store the triples in a (single) multidimensional index. In this paper, we explore this direction and study the performance of evaluating RDF queries over triples stored in a multidimensional index. We propose an efficient algorithm for processing star group pattern queries on top of a multidimensional index. By means of an extensive evaluation on real-life data sets, we demonstrate the merits of our indexing scheme and the efficiency of our algorithm.","PeriodicalId":436862,"journal":{"name":"Proceedings of the 21st Pan-Hellenic Conference on Informatics","volume":"167 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2017-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 21st Pan-Hellenic Conference on Informatics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3139367.3139408","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
Efficient management of RDF data is important due to the increasing number of data sources available in RDF, which fosters interoperability and enables data integration and interlinking. An RDF store maintains data in the form of triples, subject (S), property (P), object (O). A common approach for storing RDF data is to follow a so-called "one-triples-table" approach, where all triples are stored in a single table. For efficient data access, multiple one-dimensional clustered indexes (for example B+trees) are built, practically replicating and storing the triples in different sort orders, thus enabling efficient random access based on any subset of SPO. An issue that is largely overlooked in related research is that since triples are 3-dimensional, a natural alternative is to store the triples in a (single) multidimensional index. In this paper, we explore this direction and study the performance of evaluating RDF queries over triples stored in a multidimensional index. We propose an efficient algorithm for processing star group pattern queries on top of a multidimensional index. By means of an extensive evaluation on real-life data sets, we demonstrate the merits of our indexing scheme and the efficiency of our algorithm.