一种新型索引结构

传统Hash算法中溢出桶与主桶、溢出桶与溢出桶之间一般通过指针实现链接，对海量数据的等值查询采用指针方式效率很低。该文提出一种动态哈希索引算法，用B+树结构表示桶地址表，在桶地址表与记录键值之间建立一个B+树结构，通过二分查找可直接找到相应桶元素。实验结果表明，该算法的综合性能优于其他索引，其等值查询效率提高了15%。

Novel Index Structure

In conventional Hash algorithms, the overflow buckets are linked with not only main bucket but also overflow bucket by the pointer. It is inefficient to search a value corresponding to a overflow bucket. This paper presents a novel dynamic Hash indexing structure called DL tree which uses B+tree structure to describe bucket address table, and establishes a B+tress between the bucket address table and register key assignments. The bucket element can be found by dimidiate research. The experimental results show that its comprehensive performance is better than other indexing algorithm. It can improve 15% of the equivalent inquiry efficiency.