【摘要】 在信息爆炸的时代,全文搜索引擎必须尽快索引大量数据,以提供最佳的检索服务.本文首先分析传统独立索引的问题,以及如何通过分布式和并行索引解决这些问题.然后提出了基于映射/归约的分布式索引算法,改变了映射节点的中间数据的结构,并使用映射/归约的第二个排序函数.最后,通过实验验证了其时间和空间效率.更多还原

【Abstract】 In the era of information explosion,the full-text search engine must index vast amounts of data as soon as possible to provide the best retrieval service. This paper firstly analyzes problems of the traditional stand-alone indexing,and how to solve them by distributed and parallel indexing. Then a distributed indexing algorithm based on the map/reduce is proposed,which change the structure of map nodes’ intermediate data and uses the second sort function of Map/reduce. Finally,its time and space effectiveness is verified by experiment.更多还原