Non-uniform key distribution and address calculation sorting

• Authors: Michael E. Tarter;Richard A. Kronmal
• Venue: ACM '66 Proceedings of the 1966 21st national conference
• Year: 1966

It is currently far from uncommon for researchers to be confronted with the problem of sorting large files of data, e.g., 20,000 items or keys. One sorting procedure, known as address calculation sorting, is a two stage process in which an address calculation function is constructed by a preliminary examination of the keys. This function is used to insert each key into a location, hopefully near its position in the final sorted file. This process has been referred to as “rough focusing.” In the second phase the tentatively inserted key is compared to its neighbors to assure that it has been inserted correctly in relation to previously processed keys. This is referred to as “fine focusing.” It is currently far from uncommon for researchers to be confronted with the problem of sorting large files of data, e.g., 20,000 items or keys. One sorting procedure, known as address calculation sorting, is a two stage process in which an address calculation function is constructed by a preliminary examination of the keys. This function is used to insert each key into a location, hopefully near its position in the final sorted file. This process has been referred to as “rough focusing.” In the second phase the tentatively inserted key is compared to its neighbors to assure that it has been inserted correctly in relation to previously processed keys. This is referred to as “fine focusing.”