利用频率特征的Trie树索引快速构造算法.PDF

2013 年 4 月 第 36 卷 第 2 期 北 京 邮 电 大 学 学 报 Journal of Beijing University of Posts and Telecommunications Apr. 2013 Vol. 36 No. 2 摇 摇 文章编号:1007鄄5321(2013)02鄄0084鄄05 利用频率特征的 Trie树索引快速构造算法 张启飞1,摇 吴吉义1,2,摇 李文娟1,3,摇 吕红兵1,摇 潘雪增1 (1.浙江大学 计算机科学与技术学院, 杭州 310027; 2.杭州师范大学 电子商务与信息安全重点实验室, 杭州 310036; 3.杭州师范大学 钱江学院, 杭州 310036) 摘要: 随着物联网技术的日益成熟和云计算标准的确立以及各种智能终端的大规模出现,互联网数据呈指数增加, 为数据建立索引至关重要,为此提出一种基于词频的 Trie树索引快速构造算法,首先对索引字符串进行排序,然后 对排序文件进行预处理,预处理生成一个三元组,分别由相同字符横向偏移、纵向偏移及字符组成. 快速算法依次 扫描预处理数据的每一列,根据三元组的偏移跳过相同的字符前缀. 实验结果显示,本算法的时间明显少于传统构 造算法,优于 Aoe的双数组 Trie构造算法. 关摇 键摇 词: 索引构造; 快速算法; Trie树; 字符频率; 双数组 Trie 中图分类号: TP338郾 8摇 摇 摇 摇 文献标志码: A A Fast Trie Tree Index Construction Algorithm Using Frequency Characteristic ZHANG Qi鄄fei1,摇 WU Ji鄄yi1,2,摇 LI Wen鄄juan1,3,摇 L譈 Hong鄄bing1,摇 PAN Xue鄄zeng1 (1. College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China; 2. Key Laboratory of E鄄Business and Information Security, Hangzhou Normal University, Hangzhou 310036, China; 3. College of Qianjiang, Hangzhou Normal University, Hangzhou 310036, China) Abstract: With maturity of “the Internet of things冶 and establishment of cloud computing national stand鄄 ards, kinds of terminals appear quickly, and huge amounts of data蒺s generation exponentially increase, so it蒺s crucial to construct index for the data. A fast Trie鄄tree index construction algorithm is proposed. All the strings are sorted and then the sorted strings are preprocessed, and after that a matrix with the element of triple is generated, consisting of the character, the horizontal and vertical offset of the character. The fast algorithm scans each column in turn and skips the repeated rows and columns with the same prefix according to offset value in triple array. The experimental results show that the fast algorithm significantly reduces the construction time compared with traditional algorithm and the performance is better than Aoe蒺s double鄄array Trie construction algorithm. Key words: index construction; fast algor