To remove duplicate rows before the primary key or unique indexes are created:

 

1.

Suppose you have one identifying key filed (aKey)

This script deletes each row in the group after the first row

 

 

DELETE FROM aTable

 WHERE rowid NOT IN

 (SELECT MIN(rowid)

    FROM aTable

GROUP BY aKey);

 

 

2.

Using a correlated subquery

 

DELETE FROM aTable t1

WHERE rowid > (

    SELECT MIN(rowid)

      FROM aTable t2

     WHERE t1.aKey = t2.aKey)

 

3.

Using an analytic function

 

DELETE FROM aTable WHERE rowid IN

(

  SELECT rowid FROM

  (

SELECT rowid, ROW_NUMBER () OVER (PARTITION BY aKey ORDER BY rowid) r_no

       FROM your_table

  )

  WHERE r_no > 1

)