Skip to Main Content
R-tree and its variants didn't consider the relation between index records included in leaf nodes and storage locations in which index records are stored. Because of this reason, an index structure was designated as RS-tree in this paper. In RS-tree the index records of sequential storage locations in database files are distributed to the same leaf node or sequential leaf nodes, and the sequence of nodes of non-leaf layer is kept the same as the one of nodes of their children's layer, which leads to the reduction of consuming time effectively that the database system spends in searching records during query. We have taken a series of experiments on a data set which consisted of 2219803 records. The experimental results show that during querying successively more than 1000 records test RS-tree's speed is nearly 500 times faster than R-tree's.
Date of Conference: 19-20 Dec. 2009