Efficient SQL to calculate # of shared affiliations

Question

So I have

1. a table that stores asymmetrical connections between two persons (like a Twitter follow; not like a Facebook friend) and
2. a table that stores a person's affiliations to various groups

My task is to find, for each asymmetrical relationship, the number of affiliations shared between the "from person" and the "to person".

I made this brute force solution, but I'm wondering if brighter minds could come up with something more efficient.

select frm01.from_person_id, frm01.to_person_id, count(*) num_affl
from
(
    select lnk.from_person_id, lnk.to_person_id, ga.grp_id from_grp_id
    from links lnk
    left outer join grp_affl ga on lnk.from_person_id = ga.person_id
    group by lnk.from_person_id, lnk.to_person_id, grp_id
) frm01
inner join
(
    select lnk.from_person_id, lnk.to_person_id, ga.grp_id to_grp_id
    from links lnk
    left outer join grp_affl ga on lnk.to_person_id = ga.person_id
    group by lnk.from_person_id, lnk.to_person_id, grp_id
) to01
on  (
    frm01.from_person_id = to01.from_person_id
    and frm01.to_person_id = to01.to_person_id
    and frm01.from_grp_id = to01.to_grp_id
    )
group by frm01.from_person_id, frm01.to_person_id;

Using ANSI SQL on Netezza (which doesn't allow correlated subqueries).

TIA!

Edited to add table schema:

table lnk:
from_person_id to_person_id
1               4
2               5
3               6
4               2
5               3


table grp_affl:
person_id   grp_id
1           A
1           B
1           C
2           A
3           B
4           C
5           A
5           B
5           C
6           A

expected output:
from_person_id  to_person_id    num_affl
1               4               1
2               5               1
3               6               0
4               2               0
5               3               1

Persons 1 & 4 have 1 affiliation in common (C), 2 & 5 have A in common, 5 & 3 have B in common. 3 & 6 have nothing in common. Likewise 4 & 2.

Answer 1

You can do this with aggregation and the right joins:

select pairs.from_person, pairs.to_person, count(*)
from links pairs join
     grp_affil fromga
     on fromga.person_id = pairs.from_person join
     grp_affil toga
     on toga.person_id = pairs.to_person and
        toga.grp_id = fromga.grp_id
group by pairs.from_person, pairs.to_person;

The joins bring in the groups. The last condition only brings in matching groups between the two persons. The final group by counts them.