Abstract: Some index technologies based on graph structure improve the query speed of XML data by reducing the searching scope, but there are problems caused by irrelevant nodes comparting. This paper improves the M*(k)-index construction algorithm, and proposes a single index structure, called M**(k)-index to take place of original sequence structure. It avoids storage of nodes and edges in components from sequence, and links between components. M**(k)-index has better storage performance and higher query efficiency. Experimental results show that the improved structure has good performance.

Key words: XML data, index structure, M*(k)-index structure

摘要： 一些基于图结构的索引技术通过缩小查询范围提高对XML数据的查询速度，但存在分隔不相关节点的问题。该文改进M*(k)-index的构造算法，用单一的M**(k)-index结构取代原算法的序列结构，以避免存储序列成员中的节点与边及序列成员之间的链接，提高其存储性能和查询效率。实验结果表明，改进后的结构具有良好性能。

关键词: XML数据, 索引结构, M*(k)-index结构

CLC Number: 

• TP311