select fields, fields, ROW_NUMBER OVER (PARTITION primary key(s) order by...) seq
FROM table WHERE condition

This sql statement enables to detect duplicates entries over a key and creates a seq column which says which one is the sequence of the duplicates. This way you can take only the seq = 1 and still be able to identify all duplicates

Strategy to identify duplicates in an ETL context (for example)

sql, Oracle

PL/SQL