The UNIQUE predicate is used to determine if duplicate rows exist
    in the result table of a column select expression. Note that the
    UNIQUE predicate (in compliance with the SQL language standard)
    ignores rows with a NULL column value and ensures uniqueness for
    the other column values. Contrast this with the SINGLE predicate,
    which considers a single column value of NULL as a match for any
    other NULL value in the same column.

  (B)0unique-predicate =                                
                                                    
  qqq> UNIQUE qqq> ( col-select-expr ) qqq> 
                                                    

    If any two rows in the expression are equal to one another, the
    UNIQUE predicate evaluates to false.

    The following example determines those cities in which one and
    only one employee from the EMPLOYEES database lives.

    SQL> SELECT E.LAST_NAME, E.CITY FROM EMPLOYEES E
    cont> WHERE UNIQUE
    cont> (SELECT * FROM EMPLOYEES EMP
    cont> WHERE EMP.CITY=E.CITY);
     LAST_NAME        CITY
     Harrison         Boston
     Smith            Bristol
     McElroy          Cambridge
     Kilpatrick       Marlow
     Sciacca          Munsonville
     Vormelker        Rochester
     Dement           Sanbornton
     Babbin           Sanbornville
     Keisling         Twin Mountain
     Ziemke           Winnisquam
     Johnston         Wolfeboro
    11 rows selected