Minimum cost selection of secondary indexes for formatted files
ACM Transactions on Database Systems (TODS)
Analysis and performance of inverted data base structures
Communications of the ACM
A relational model of data for large shared data banks
Communications of the ACM
| Hi-index | 0.00 |
In this paper we define a tree-structure, the reduced cover-tree (RCT), for implementing both classical indexes and multiattribute search structures and we outline its application in the data access model of the relational multi-processor data base machine SABRE. The kernel of this data access model is an RCT describing the placement of a relation that could be either a multi-attribute clustering scheme also used to - in some extent - substitute secondary indexes, or a non-dense unique key index. If the placement is achieved by multi-attribute clustering a unique key index can be nicely constructed with the same tree-structure. The RCT is represented by a bit-table in a most economic way and thus the number of accesses to memory when retrieval can be considerably cut down. On the average the memory occupation rate should turn about 75 per cent.