Processing Inequality Queries

Bernstein and Goodman showed that natural inequality ( NI) queries can be processed efficiently by semijoins, if there are no multiple inequality join edges, nor cycles with one or zero doublet. In this paper procedures to hand1 e these cases efficiently are given. Multiple inequality join edges can be processed by multi-attribute inequality semijoins. Two procedures based on generalized semi-joins for cyclic NI queries (with one or zero doublet) are developed. Semi-join is a useful operation to reduce the processing cost in distributed databases and database machines [BERNCBlOl] [BERNG81111. Its processing power, however, is limited because not all queries can be solved using semi-joins only. When queries consisting of natural joins of relations (called NJ (Natural Join) queries) are considered, queries in the class called tree queries can be solved using semi-joins only but the rest of queries (called cyclic queries) cannot [BERNG81111. We have introduced generalized semi-joins and developed procedures for cyclic queries using generalized semi-joins [KAMBY8206]. We have also developed several methods to transform cyclic queries into tree ones utilizing data dependencies IKAMBY83051. Using these processing methods, cyclic queries also can be solved Permission to copy without fee all or part of this moteriol i( granted provided that the copies ore not made or distributed for direct commerdal advantage, tht VLDB copyright notice and the title of the publication and its date appear, and notice is given that copying is by permission of the Very Large Data Base Endowment. To copy otherwise, or to republish, requires a fee and/or special permission from the Endowment. Proceedings of the Tenth International Conference on Very Large Data Bases. BASED ON GENERALIZED SEMI-JOINS