Research On Query Rewriting Of Distributed Knowledge Graph

In recent years,with the development of Internet information technology,the amount of data generated by the Internet has increased rapidly.These large-scale data generally have the characteristics of low value density,rich data types and extensive data sources.Therefore,when the user carries out information retrieval,it is more desirable to obtain the knowledge contained in the page than to obtain the title information of the page with the user's retrieval conditions.Ontology-based RDF is the focus of current research,and SPARQL,a knowledge graph query language for ontology RDF data,has also been studied and applied.However,due to the lack of reasoning ability,the traditional SPARQL query statement can not find the implied semantics in the ontology RDF data,resulting in incomplete query results and unable to fully obtain the information users want.At the same time,with the development of big data technology,the scale of RDF data is gradually increasing.In order to meet the requirements of large-scale data storage and efficient and real-time query,how to reasonably store and query ontology RDF data has become the focus of current research.Due to the weak scalability of previous relational databases,it has been unable to adapt to large-scale data storage and query.However,some existing distributed data storage query methods have improved the query efficiency to some extent,but they do not utilize the semantics between ontology RDF data.Therefore,in order to let the user's query result is more perfect and improve the efficiency of query,this paper set out from the SPARQL query statements,more appropriate use of ontology semantic reasoning datalog logic program to rewrite the traditional SPARQL statement,fully excavating implicit connection between RDF data model,let SPARQL language reasoning ability,realizes the query results and the query matching of semantic level,so that the query result is more perfect;At the same time,this paper adopts distributed storage and query methods,and improves query efficiency by optimizing SPARQL query statements on the premise of ensuring the completeness of query results.The specific work of this paper is as follows:(1)According to the correlation between SPARQL and DATALOG statements,a conversion method between SPARQL and DATALOG language is given,and then rewritten for DATALOG query language,so as to realize the rewriting of SPARQL query statements indirectly,thus realizing the completeness of query results of SPARQL statements;(2)Incomplete database is given to the distributed database migration scheme,first of all,according to the SPARQL query mode characteristic,design special HBase table for SPARQL query,and then presents a about RDF data conversion method,convert RDF data to the structured format suitable for storage to HBase,finally through graphs computing framework to RDF data mass storage to HBase;(3)According to the proposed ontology RDF data storage method,according to the characteristics of SPARQL query statement,by analyzing the relationship between variables in SPARQL statement and corresponding query statement,the optimization algorithm of SPARQL query statement is proposed.In the final experiment stage of this paper,LUBM data set was used as a comparison test.First,the query results of the rewritten statement and the original statement were tested and compared to verify that the rewrite effect of the query statement was indeed extended on the basis of the original query statement.Then through the experiment of distributed environment and stand-alone environment,as well as the comparison of query strategies before and after optimization in distributed environment,the distributed query optimization method in this paper is verified to have a certain high efficiency.