Efficient Search of Multi-Dimensional B-Trees

Data in relational databases is frequently stored and retrieved using B-Trees. In &cis,ion isugprt applications the key of the B-Tree frequently involves the concatenation of several fields of the relationdl’ table. During retrieval, it is desirable to be able to access a small subset of the table based’ on partial key information, where some fields of the key may either not be present, involve ranges, or lists ‘of values. It is also advantageous to altow. this type, of access-with gen&il expressions involving any combination of disjuncts on key columns. This paper &scribes a method whereby BTrees can be eficiently used to retrieve small subsets, thus avoiding large scans of potentially huge tables. Another benefit is the ability of this method to reduce the need for additional secondary indexes, thus saving space, maintenance cost, and random accesses.