Removing duplicates in child tables

Gurus,

I have a case, where three tables are used as source and it was JOINED. From two child tables, I want to remove duplicates on few keys.

To remove duplicates, either

-->I have to put SQL transform with group by clause before JOIN transform

or

-->Need to put order by and gen_rownum_by_group and have to filter the record with rownum=1 before JOIN transform

Both ways are not pushing down the SQL to Oracle. Is there is any other better approach to handle this? Kindly guide me.

Thanks