The ability to use the target table of a DML statement more than once.

Delete duplicates from a table:

DELETE FROM some_table
WHERE (a,b) IN (SELECT a,b
                FROM some_table
                GROUP BY a,b
                HAVING count(*) > 1);

Swap values between rows:

UPDATE some_table
   SET name = CASE id
                 WHEN 1 THEN (SELECT name FROM some_table WHERE id = 2)
                 WHEN 2 THEN (SELECT name FROM some_table WHERE id = 1)
              END
WHERE id IN (1,2);

Back to the SQL Feature Comparison