Interactive applications such as expert systems, CAD/CAM and multimedia impose an increasing demand on a data management system that efficiently supports basic operations on very large files and provides data retrieval with a guaranteed single disk access. The synergism of a conventional B+ tree and a hash function represents a possible solution to the problem. We have developed a class of algorithms that allow a single disk access. The purpose of the paper is to compare and contrast several fast and simple hash functions suggested in the literature that can be used in such a class of algorithms.