This is now solved. My data was in a SAS dataset, and the fix below worked. Used proc sql for this. proc sql; Create table dataset as id, amt1 format comma11.2, amt2 format comma11.2 from dataset; Quit; Result: 2345.56 and 87.60 from original $87.60,$2345.56 In this case the original source had no bearing, as it ended up as a SAS Dataset with $Dollar14.2 and Informat 14.
... View more