How to get count of two columns of a table based on PK of another table?

Question

I have two tables defined as:

CREATE TABLE airports (
  code char(3) not null primary key,
  name varchar(100) not null,
  city varchar(50) not null,
  state varchar(5),
  country char(2) not null references fcountries(code)
)

and

CREATE TABLE flights (
  departure char(3) not null references airports(code), -- airport code
  arrival char(3) not null references airports(code), -- airport code
  dep_time time not null,
  airline char(2) not null references airlines(code)
)

I need to get the number of flights departing from and arriving at an airport using postgres. So, I wrote

select
  code,
  count(departure)
from airports 
join flights 
 on flights.departure = airports.code group by code;

and for departure similarily,

select
  code,
  count(arrival) as Arrival
from airports 
join flights 
  on flights.arrival = airports.code group by code;

But I need to combine these two and get the arrival-departure count in the same result. I am not sure how to do it? I wrote this query

select
  code,
  count(departure),
  x.arrival
from (
  select count(arrival) as arrival
  from airports 
  join flights 
    on flights.arrival = airports.code group by code
) x, 
airports 
join flights on airports.code = flights.departure
group by code, x.arrival
order by code;

But the result is not correct as 'code' is repeated in this result. I am pretty new to sql and not sure how to write this.

Answer 1

The simplest method is a lateral join and group by :

select v.code, count(*) as total,
       sum(is_departure) as num_departures,
       sum(is_arrival) as num_arrivals
from flights cross join lateral
     (values (departure, 1, 0), (arrival, 0, 1)
     ) as v(code, is_departure, is_rrival)
     on f.departure = a.code
group by v.code;

Note that the JOIN to airports is unnecessary, unless you actually want to bring in more information that you haven't mentioned in the question.

Answer 2

You could use the two query as subquery and join

    select t1.code,  t1.count_departure, t2.count_arrival 
    from (
        select code, count(departure)  count_departure
        from airports join flights on flights.departure = airports.code 
        group by code

    ) t1  
    inner join  (
        select code, count(arrival) as  count_arrival 
        from airports join flights on flights.arrival = airports.code 
        roup by code

    ) t2 on t1.code  = t2.code ;

Answer 3

I think alias t2 cannot be used like that after join that's why you are getting the error. Try this:

select t1.code, t1.count_departure, t2.count_arrival from ( select code, count(departure) count_departure from airports join flights on flights.departure = airports.code group by code ) t1 join ( select code, count(arrival) as count_arrival from airports join flights on flights.arrival = airports.code group by code ) t2 on t1.code = t2.code;

Answer 4

A quite simple method is just to use a correlated COUNT subquery in the SELECT clause:

select a.code,
    (select count(*) from flights f where f.departure = a.code) as departures,
    (select count(*) from flights f where f.arrival = a.code) as arrivals
from airports a;