Finding the shortest path in a graph is a fundamental operation that allows complex relationships discovering in a graph database.This paper considers top-

k

shortest-path queries on RDF(Resource Description Framework) graphs.To solve this problem

we provide a component-based index for path queries and present an approach to answer the top-

k

shortest-path query.To efficiently process queries

we propose frequent path strategy and structural pruning.For frequent path strategy

we precompute the top-

k

shortest-paths between frequent entities obtained from the frequent paths.For the structural pruning

we find out the termination condition for queries by considering the structural information of component-based index and derive two lemmas to guide this pruning mechanism.Experiments results show that the qu

ery runtime is improved by 92% using our answering approach and the construction time of our index for RDF graphs is speeded up by 56%.Meanwhile

the size of this index is only 0~0.2 times larger than that of the raw graphs.