Achieving effective and efficient attributed graph data management using lucene
Jiaxin Zou, Bo Lang, Jiheng Zhao, Yishuai Zhao · 2018
How to manage graph data which have complex graph structure reasonably and efficiently is a big challenge. Graph database becomes a new choice due to its natural graph processing ability. Nevertheless, it is still in the development stage with some problems that need to be solved. For instance, the formal definition of attributed graph model, the query efficiency and effectiveness remain to be perfected. In this paper, we make a formal definition of the attributed graph model, and implement an attributed graph database MyGraphDB based on Lucene which is a high-performance, full-featured text search engine library. The introduction of Lucene not only brings the improvement of query efficiency, but also takes advantage of the abundant attribute information of the nodes and edges. We compare our MyGraphDB with another two graph databases SparkSee and Neo4j. The results show that MyGraphDB has obvious advantages in terms of efficiency and the diversity of operations.