The LSD tree: spatial access to multidimensional and non-point objects
VLDB '89 Proceedings of the 15th international conference on Very large data bases
The R*-tree: an efficient and robust access method for points and rectangles
SIGMOD '90 Proceedings of the 1990 ACM SIGMOD international conference on Management of data
Access method concurrency with recovery
SIGMOD '92 Proceedings of the 1992 ACM SIGMOD international conference on Management of data
Multidimensional access methods
ACM Computing Surveys (CSUR)
The K-D-B-tree: a search structure for large multidimensional dynamic indexes
SIGMOD '81 Proceedings of the 1981 ACM SIGMOD international conference on Management of data
R-trees: a dynamic index structure for spatial searching
SIGMOD '84 Proceedings of the 1984 ACM SIGMOD international conference on Management of data
A Robust Multi-Attribute Search Structure
Proceedings of the Fifth International Conference on Data Engineering
The LSDh-Tree: An Access Structure for Feature Vectors
ICDE '98 Proceedings of the Fourteenth International Conference on Data Engineering
The X-tree: An Index Structure for High-Dimensional Data
VLDB '96 Proceedings of the 22th International Conference on Very Large Data Bases
Improving the R*-tree with outlier handling techniques
Proceedings of the 13th annual ACM international workshop on Geographic information systems
Multidimensional Binary Search Trees in Database Applications
IEEE Transactions on Software Engineering
Improving performance with bulk-inserts in Oracle R-trees
VLDB '03 Proceedings of the 29th international conference on Very large data bases - Volume 29
| Hi-index | 0.00 |
The R-tree [7] family is the most popular multi-dimensional index method. The R-tree, however, has overlaps among index entries and its index page fanout decreases rapidly as data dimension increases. Furthermore, the R-tree has poor concurrency performance. For frequent-update multi-dimensional point data sets, the hB-pi [5] tree is a better choice than the R*-tree. But the hB-pi tree (and all other kd-tree based access methods) indexes the whole space no matter whether or not there is any data in some sub-spaces. Indexing empty space(i.e., space without data inside) leads to unnecessary data page accesses which increase with growing dimension. This paper addresses this problem by proposing the hB-pi* tree, which efficiently indicates empty spaces and improves range query performances while preserving the hB-pi's high fan-out and good concurrency. Our methods can be applied to any kd-tree based access methods, and our claims are supported by extensive experimental evaluation.