SQL按多列中的值对分组并消除相同的元组

Question

I have a database table containing flights. (The origin/destinations are in IATA airport codes.)

id   origin   destination
1    AMS      BON
2    BON      AMS
3    EIN      GDN
4    GDN      EIN
5    EIN      GDN
6    AMS      AGP

I am trying to find the number of the flights between each two airports, irrespective of the direction of the flight. So I am looking for this result:

origin    destination    count
AMS       BON            2
EIN       GDN            3
AMS       AGP            1

What I have so far is this query:

SELECT c.origin, c.destination, count(c.origin) as count FROM 
(
SELECT a.origin as origin, a.destination as destination FROM Flights a 
UNION ALL SELECT b.destination as origin, b.origin as destination FROM Flights b 
) c 
GROUP BY origin, destination;

Which gives the following result:

origin    destination    count
AMS       BON            2
BON       AMS            2
EIN       GDN            3
GDN       EIN            3
AMS       AGP            1

So the problem is that there are duplicate tuples. How do I eliminate those?

It seems similar to this question, but not similar enough I'm afraid. Eliminate tuples with reverse relation and no primary Key

Answer 1

An easier approach is to use least() and greatest() :

SELECT least(origin, destination) as city1,
       greatest(origin, destination) as city2, count(*) as cnt
FROM Flights f
GROUP BY least(origin, destination), greatest(origin, destination)