As more libraries automate for the First time or migrate to more sophisticated integrated systems in a time of fiscal restraint and, in many cases, down‐sized staffing, the ability to automate as many functions as possible (particularly tedious ones, such as database cleanup) would seem desirable. The project described in this article is a good example. If the microcomputer program for cleaning up Kent State University's NOTIS database was not written, the project, which probably should have begun years ago, would still not have been implemented. The solution presented here addresses these key issues realistically, within the confines of the programming resources available within many libraries.