SAS tip: check row counts before trusting a join

State the relationship

Write down whether the join should be one-to-one, one-to-many or many-to-many. Profile duplicates on both sides before joining. A repeated key on each side can multiply records, changing counts and downstream summaries while producing valid SQL.

Reconcile the result

Compare distinct subject counts, row counts and unmatched keys before and after the join. Retain unmatched records for investigation rather than silently discarding them. A DATA step merge and a SQL join do not have interchangeable behaviour for every duplicated-key scenario.

Sources & further reading

• SAS — Documentation

Sources checked 20 September 2026. Implementation commentary reflects a practical review perspective; source material may change.