It works. Thanks thanks for the help Reeza! I still have a problem on the actual data, because I still have duplicates. So for example, I have John Doe Accounting John Doe Accounting Duplicates haven't been removed, for example, using dedup under proc freq. This seems to be because the underlying values are formatted. So Accounting was Accounting_secretary and Accounting_executive. Basically, John might have multiple roles under accounting. A format is applied to reduce them to accounting, and the resulting duplicates are not getting removed.
... View more