摘要
新的利用连接索引的算法Jive,它用中间产生的临时文件和分割输出连接结果的方法,采用已有的数据结构-连接索引,只需要对输入关系的一次扫描,即可完成连接.在一般情况下优于Valduriez(1987)所提出的连接算法,在输入关系很大的情况下性能尤为突出。
By storing the intermediate results as temporary files and outputting the final join results in partitions,the new join algorithm Jive which uses the existed data struct join index ,requires only one pass through each input relation to complete the join.In most situations,it has better performance than the algorithm proposed by Valduriez（1987）,and it performs best when the input relation is very large.
出处
《计算机工程与应用》
CSCD
北大核心
2006年第12期180-182,共3页
Computer Engineering and Applications
