Using common subexpressions to optimize multiple queries

Abstract

The problem of identifying common subexpressions and using them in the simultaneous optimization of multiple queries is dealt with. In particular, emphasis is placed on the strategy of selecting access plans for single queries and their integration into a global access plan that takes advantage of common tasks. A dynamic programming algorithm is presented for the selection of individual access plans such that the resulting global access plan is of minimum processing cost. The computational complexity of this algorithm represents a significant improvement over existing algorithms.<>