The table f has these values:
name1 name2 count
Harrison Ford Mullah Omar 3
Harrison Ford Michael Caine 6
Michael Caine Mullah Omar 2
Michael Caine Harrison Ford 6
Mullah Omar Michael Caine 2
Mullah Omar Harrison Ford 3
How to choose only the distinct pairs?
select
distinct name1,
distinct name2, count from f group by name1
Given that you have a duplicate for each pair, you can just use:
select f.*
from f
where name1 < name2;
If you did not have the reverse in all cases, you can use:
select f.*
from f
where f.name1 < f.name2 or
not exists (select 1
from f f2
where f2.name1 = f.name2 and f2.name2 = f.name1
);