This has me beat. I have 2 tables, address_detail and address_locality.
fields in address_detail include a PKey = address_detail_uid (int), 
locality_uid (int), address_postcode (character(4)) and other stuff.

address_locality including a PKey = locality_uid (integer).

There is a foreign key in address_detail linked --> locality_uid.

There are no address_detail rows with a NULL locality_uid, so each 
locality in address_detail has a matching locality record.

My task is to transfer postcode from address detail to locality. There 
are many address details with the same locality and postcode.

Step 1: I get the number of distinct localities in address_detail.

SELECT DISTINCT locality_uid FROM address_detail;

15,195 localities

Step 2: I get the number of localities in the locality table

SELECT locality_uid from locality;

16,398 localities

OK, so there are a number of localities which don't have addresses. 
That's OK because there are a lot of parks, fields, beaches etc etc in 
there as localities.

Step 3: I now get the number of distinct locality/postcode
         combinations in the address table, to see if the total
         localities match.

SELECT DISTINCT locality_uid, address_postcode FROM address_detail;

17,824 rows

So we have a number of localities with more than 1 postcode. Now I want 
to determine which ones.

Step 4: Check for duplicate postcodes.

SELECT foo.address_postcode, foo.locality_uid, COUNT(1) as COUNT
FROM
    (SELECT DISTINCT address_detail.address_postcode,
                     address_detail.locality_uid
     FROM address_detail) AS foo
GROUP BY foo.address_postcode, foo.locality_uid
HAVING COUNT(1) > 1
;

0 Rows

Can someone tell me where are the duplicates?