CHECKR: an efficient table-driven facility for input record data validation

Abstract

It is desirable to detect and to identify properly any input data errors in a data collection system. The most common errors include hardware errors, field range errors, incorrect character-in-field errors, and records-out-of-sequence errors. In order to process these error conditions in a general manner, a set of Xerox FORTRAN and assembly language routines has been developed. These routines, collectively referred to as the CHECKR system, are table-driven and, therefore, can be used to validate input data from many different sources. A number of tables are created which specify the allowable conditions that can exist, and CHECKR uses these tables to produce an easy-to-read error message report. CHECKR has been used successfully to validate data for the Automated Data Repository (ADR) Project, a system to collect and process information concerning exercise and its relation to the prevention of coronary heart disease.