Index And Query Technologies Study For Large RDF Datasets Based On Signature Graph Theory

The RDF data and the SPARQL query have a graph structure, so the RDF data storage and query based on graph model has become a way to intuitively and effectively. At this time the query process become a subgraph matching process which is very time-consuming, so indexing technology is the key to solve this problem. In existing schemes, the excessive matching causes unnecessary time consumption. In order to solve this problem, this paper proposes a new query algorithm, and provides a solution of index maintenance, effectively solved the problem. The main work of this paper as follows:(1) Aiming at the limitation of map data storage scheme of the original large quantity, ergodic and find difficult, this paper puts forward two kinds of storage scheme, which makes the amount of stored data is greatly reduced, and the traversal and search efficiency is not significantly reduced, in some special cases, such as the query variable is much better.The original eigenvalue distance measurement and index node splitting method is not reasonable cause node splitting operation frequent. This paper presents a new eigenvalue distance measurement and index node splitting algorithm, which can effectively reduce the node splitting operation, reduce the index construction time.(2) The original query algorithm repeatedly query times more, this paper designs a new query algorithm, which can consume less memory and reduces the number of queries, query efficiency improvement. This paper discusses the causes of this phenomenon. The original maintenance technology for RDF indexes does not put forward the concept node merging, so the excess data delete will destroy the index structure of the problem. This paper describes the design of a merge node, which can maintain the balance of the index structure, reduce the index node redundancy.