Then there's the question of "testing". If attempting this on 14m records, it will SORT them all before the SUM fails (and you may have got unlucky with the previous values in the "filler").
To see what the OVERLAY was doing, take out the SORT and the SUM and run on a test file (even the full file, with OPTION STOPAFT=10 or the like). When getting used to SORT Control Cards, I'd do add a bit at a time. When looking for problems, take a bit off at a time. Either way you get to "see" what you have said should happen to the data. 219 bytes for the key? 14m+ records? 11-digit number for duplicate keys? What is the significance of the duplicate keys? Is the data not already in that sequence at some point (else there would seem to be little significance)? ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
