Abstract
We examine the efficiency of generalized hash-coding algorithms for performing partial-match searches of a random-access file of binary words. A precise characterization is given of those hash functions which minimize the average number of buckets examined for a search; and a new class of combinatorial designs is introduced which permits the construction of hash functions with worst-case behavior approaching the best achievable average behavior in many cases.

This publication has 4 references indexed in Scilit: