Research On Storage And Query Technology Of RDF Data Based On Graph

Resource Description Framework?RDF?,as an emerging data representation,provides a semi-structured way to describe resources for applications.The wide application in the fields of semantic Web,social network,digital library and bioinformatics has made the management of RDF data become one of the hotspots in academic and industrial research.This paper studies and analyzes the problems of excessive storage space and low query efficiency of RDF data.The main work is as follows:?1?Aiming at the problem that the storage space of large-scale RDF data is too large,a representation method of k²_PP-MDD-triples is proposed based on k²_PP-tree and multi-valued decision diagram technology.According to the predicate,the data set is vertically divided into disjoint subsets of?subject,object?,then the subject and object are binary coded.According to the coding of subject and object,the triple is coded.The k²_PP-MDD-triples are constructed by the set of triple codes.In the process of constructing k²_PP-MDD-triples based on triple encoding set,redundant nodes such as null values and isomorphic subtrees in RDF data storage based on k²_PP-tree are merged and deleted by MDD simplification rules,which reduces the number of nodes and makes the RDF data storage structure more compact.?2?This paper studies the relationship between RDF data management and MDD logical operations,transforms RDF data management into logical operations of multi-valued decision diagram.And we provide the implementation method of triples pattern query,basic graph pattern query and adding or deleting RDF triples under k²_PP-MDD-triples representation.Experiments on DBTune classical data,DBpedia Mappings Wiki data sets and LUBM-2 data sets show that the k²_PP-MDD-triples method only stores56.78%to 85.30%of the nodes in the k²_PP-tree method.?3?This paper studies the relationship between RDF triple and graph model.And transform RDF data query into sub-graph matching on large graph.Then analyse the existing RDF data query methods based on graph,and propose a new RDF query algorithm RDF-SM.By designing a new vertex sorting method and combining four filtering constraints designed for vertex attributes and edge tags,the algorithm achieves fast query.The effectiveness of the algorithm is verified by comparing the query efficiency of the triplet pattern query and the basic graph pattern query with MonetDB on the LUBM-2dataset.