A HOW on Hashing

Russ Lavery


Hash tables are underutilized, and very fast, techniques developed for table lookup – that now do much more. A key feature of has use is the avoidance of CPU and RAM intensive sorting of data during table “merges”. Understanding the behavior, and even the “shape”, of this memory resident object can be a bit of a challenge to the typical SAS programmer. This paper, with its cartoon format, endeavors to make this new programming technique more approachable.

Hash tables use a new syntax format that is similar to the object oriented syntax of other languages. Hash tables are created in a data step and exist only as long as that data step is running. It exists in RAM and are separate from the program data vector.

It should be remembered that SAS has hash tables due to the work of Dr. Paul Dorfman and the responsiveness of the people at SAS to user input.