组中每个子组的最大值

Question

So I have a table with following data :

X------------------Y
customer1------A
customer1------A
customer1------B
customer2------B
customer2------B
customer2------B
customer2------C
customer2------C
customer2------C
customer2------C
customer2------C

So I need to find what max(Y) does each customer have. So customer1 must have B and customer2 must have C (I'm using postgresql)

Answer 1

You want the "mode" in statistics. Here is one method to get exactly one value per x (ignoring duplicates):

select distinct on (x) x, y
from (select x, y, count(*) as cnt
      from t
      group by x, y
     ) xy
order by x, cnt desc;