To get a list of duplicate records, you can join the table on itself. Something like this ought to do it:

SELECT a.unique_id 
  FROM my_table as a, my_table as b 
 WHERE a.field1 = b.field1
   and a.field2 = b.field2
   and a.field3 = b.field3
   and a.unique_id != b.unique_id
[download]

This will return a list of all unique_id's where two rows share the same value of field1, field2, and field3. Note that it will return the unique_id of Both matching rows, not just one of them.

I don't have a suitable dataset to experiment on, but you might be able to also do a group by the fields which must match, and select max(unique_id) instead, to grab one row ID from each group of matching stuff. You might still need to use perl to loop through and delete results, but at least you'll then be doing a linear operaton instead of exponential in perl.

Update: changed unique_id to a.unique_id

You can group by a.unique_id to make sure you only get one row per unique id. However, in the case where you know that a large portion of your rows have duplicates, you're right: this still doesn't give you any information you don't already have Slightly better would be select a.unique_id, b.unique_id so you at least know which row each row matches. Still less than ideal, I agree.

Alan