Record duplicati in postgresql

create table t2 as
select distinct on (col1, col2, col3, col4) *
from t;

drop table t;
alter table t2 rename to t;