Hello John, how about using 2 separate DBOPEN's for the 2 datasets to process. I did not tried it, but in the past I used this technique in similar situations. Regards, Guenter Kuhn Dortmund, Germany