Joseph,

For starters, trap any possible divide by zero errors if numerics are calculated in a program. Otherwise, track down the source of overflows by working backwards to the source. What maximum range is permitted? Trap data entry values. Do file structures that share data have the same numeric precision? Are there any string to numeric conversions involved that won't work right? Are formatting descriptions responsible?

HTH,
Allen

>Is there a way to trap for numeric data overflows? I'm SCATTERing and GATHERing between VFP tables, and I'm also uploading to SQL Server. A numeric overflow prevents my programs from working. How can I look for those kinds of problems and set them aside so my programs will continue working?