摘要： 针对文件中存在字符属性和数值属性特点的数据，提出一种基于B+树的2层混合索引结构。在索引创建过程中，将文件中的数据根据其类型进行分类，第1层是为数据属性建立B+树索引结构，第2层是根据不同的数据类型建立不同的索引结构。实验结果表明，该索引结构能实现跨数据范围的检索，提高索引的创建速度和空间利用率。

关键词: 倒排索引, B+树, 混合索引, 时间开销, 空间利用率, 查询效率

Abstract: Considering the numerical attribute data’s cross-data range retrieval needs for the files that have the property of character and number, a 2-layer hybrid index structure based on B+ tree is proposed. That data in the files is classified according to their types in the process of creating index, the first layer sets up a B+ tree index structure for the attribute of the data and the second layer constructs different index structure according to the different types of data. Experimental results show that the hybrid index structure not only adapts to the above retrieval needs, but also effectively improves the index creation speed and the space utilization.

Key words: inverted index, B+ tree, hybrid index, time overhead, space utilization, query efficiency

中图分类号: 

• TP311