Unified SPARQL Query-analyzing Language Design And Theoretical Research

It is necessary to frequently use graph querying and graph analytics when solving complex,large-scale graph processing tasks.However,graph query and graph analytics are not unified yet in theory,resulting in the reliability of the results of querying and analytics cannot be guaranteed,which leads to the limits of optimization of querying and analytics.Therefore,how to unify graph querying and graph analytics,then design a unified theoretical framework,including graph querying and graph analytics is very important for solving large-scale graph processing tasks efficiently.Based on SPARQL,we establish a unified theoretical basis of graph querying and graph analytics,including its syntax and semantics,and study its basic theoretical properties,namely expressiveness and complexity.We introduce lots of graph analytical algorithms and construct the analyzing clauses,then propose SPARQL Analyze Query language.For different graph processing problems,the language has two manifestations,which are called single-layer SPARQL Analyze Query language and nested SPARQL Analyze Query language,respectively.Nested SPARQL Analyze Query language has two forms which are inner-outer nested and peer nested.Based on the original SPARQL semantics,we construct the new semantics,then prove the correctness of the two nested forms.We prove the expressive problem between SPARQL Analyze Query language and SPARQL four query forms,etc.And analyze the complexity of SPARQL Analyze Query language.Finally,the simplicity of SPARQL Analyze Query language is verified by comparison experiments.The language we propose supports graph querying and graph analytics on weighted RDF graphs,and provides the theoretical basis for graph querying and graph analysis in the industry,then guarantee the reliability of the results and the optimization of graph querying and graph analytics.Finally,we provide a new solution for unifying graph querying and graph analytics.