>Hi Tim ----
>
>>
>>We are currently using the first appoach of indexing (STR(GROSS,12,2)+INVOICE) and storing the first record of each in an array. We then start with the smallest value and check the array for a match. If a match is found, SELECTing the work area, SCATTER, SELECTing the resulting table, GATHER. If no match, increment the array variable for the current work area to the current record value, etc. Clear as mud? ;->
>
>Wow....you said it (clear as mud) :-)
>
>Ummm...so, is it safe to say that you are storing *one* instance of a duplicated row in your results table, unchanged? You are not summing the dupes or care where they came from?

Don't care what table they came from. There is some other information in the record that helps me to determine where it would have originated. Storing all rows that have duplicates. Reason is that other information in the records would indicate if the records are true dups or maybe correct entries.