SQL & PL/SQL :: Merge Vs Cursor Processing

My requirement is ..

SQL> CREATE TABLE EMP (EMPID VARCHAR2(100), U_EMPID VARCHAR2(100));

Table created.

SQL> INSERT INTO EMP VALUES ('DUMMY1','J0001');

1 row created.

SQL> INSERT INTO EMP VALUES ('DUMMY2','J1002');

1 row created.

SQL> INSERT INTO EMP VALUES ('DUMMY3','J1003');

1 row created.

SQL> SELECT * FROM EMP;

EMPID U_EMPID
------------------
DUMMY1 J0001
DUMMY2 J1002
DUMMY3 J1003

in this example empid is dummy empid (which will be issued for contractors), once the contractors are considered to be taken into payroll that u_empid will be updated by HR manually, we need to replace empid with u_empid. then data will become like this.

SQL> SELECT * FROM EMP;

EMPID U_EMPID
---------------------------
DUMMY1 J0001
DUMMY2 J1002
DUMMY3 J1003
J1001 NULL
J1002 NULL
J1003 NULL

now my requirement is there are approximately 60 child tables are there for EMP table, and EMPID is the foreign key for all of them with "on delete cascade" defined on those. Now first i need to create records in all the child tables with actual employee id's (which is u_empid). if the actual id is already exists then just need to update rest of the columns in that table except the EMPID.

Once the records are created with actual empid, then i will issue a delete command to remove dummy employee id's in emp table which will take care of deletion in child tables as well.

I was thinking of two options, one is go with Merge other is write a cursor and handle the update in exception block using when duplicate value on index feature.