Abstract: It’s a new way of solving the large amount of RDF(Resource Description Framework) query problem to use distributed technique to realize the SPARQL(Simple Protocol and RDF Query Language) Query．At present,most of the RDF queries based on Hadoop have to use multiple MapReduce jobs to complete the task,resulting in waste of time．In order to overcome this drawback,this paper proposed MRQJ(using MapReduce to the query and the join) algorithm to perform distributed SPARQL query．The algorithm can be divided into join plan generation and SPARQL query execution two parts:join plan generation uses greedy strategy to generate the most optimal join scheme,and only one MapReduce job should be done to get the query results in SPARQL query execution．The experiment on the LUBM test data set was made.The experimental results show that MRQJ method query efficiency is higher when the case of a query is more complicated.

Key words: RDF,Hadoop,SPARQL query,MapReduce