计算和求和值sql

Question

I have two tables: BORDERS with country1,county2 ,COUNTRYPOPS with country,year and population, and I want for each country, list its population and the total population of all neighboring countries.(and if it has no neighboring countries, NULL) so far I can find the max population for each country and how many neighbors each country has, but I cannot find a way to sum up the population of all neighbor countries. Any Ideas?

SELECT country, MAX(population) FROM COUNTRYPOPS GROUP BY 1 ORDER BY 1
---------------
SELECT countries, COUNT(countries) as have_borders
FROM
( SELECT country1 AS countries
FROM borders
UNION ALL 
SELECT country2
FROM borders ) t
GROUP BY countries
ORDER BY countries;

Answer 1

How about something like

SELECT country, pop, SUM(cc.population) AS neighboursPop
FROM (
   SELECT c.country AS country, c.population AS pop, b.country2 AS neighbour
   FROM countrypops c
       LEFT JOIN borders b ON b.country1 = c.country
   ) t 
       LEFT JOIN countrypops cc ON cc.country = t.neighbour
GROUP BY country;

Didn't test it, but the idea is that.

UPD: or it could even be flattened like

SELECT country, pop, SUM(neighPop) AS neighboursPop
FROM (
   SELECT 
       c.country AS country, 
       c.population AS pop, 
       b.country2 AS neighbour, 
       cc.population AS neighPop
   FROM countrypops c
       LEFT JOIN (borders b 
                    LEFT JOIN countrypops cc ON cc.country = b.country2)
       ON b.country1 = c.country
   ) t 
GROUP BY country;

Answer 2

I think a correlated subquery is the easiest way:

select cp.country, cp.pop,
       (select sum(cp2.pop)
        from borders b join
             countrypops cp2
             on b.country2 = cp2.country
        where b.country1 = cp.country
       ) as neighboring_pop
from countrypops cp;

This does not have an aggregation in the outer query, so it should be more efficient than an approach that does.