Update table column based on multi row comparison operation in Postgres

rshar :

I have following table in postgres

id          num
EBI-1002144 1
EBI-1002144 1
EBI-1002142 2
EBI-1002142 2
EBI-1002635 1
EBI-1002635 1
EBI-1002635 2
EBI-1002635 2
EBI-1003351 1
EBI-1003351 1
EBI-1003351 2
EBI-1003351 2
EBI-1003469 1
EBI-1003469 1
EBI-1003469 2
EBI-1003469 2
EBI-1003574 1
EBI-1003574 1
EBI-1003574 2
EBI-1003574 2

I want to append another column to this table based on following conditions:

--> group by id 
--> calculate max of num
--> if the value of num > 1 per id, then assign the id as label 'A' else 'B'

I am able to find out the max value but unable to figure out how to assign the value to each row with common id.

The expected output is:

id          num  label
EBI-1002144 1    B
EBI-1002144 1    B
EBI-1002142 1    A
EBI-1002142 2    A
EBI-1002635 1    A
EBI-1002635 1    A
EBI-1002635 2    A
EBI-1002635 2    A
EBI-1003351 1    A
EBI-1003351 1    A
EBI-1003351 2    A
EBI-1003351 2    A
EBI-1003469 1    A
EBI-1003469 1    A
EBI-1003469 2    A
EBI-1003469 2    A
EBI-1003574 1    A
EBI-1003574 1    A
EBI-1003574 2    A
EBI-1003574 2    A

Gordon Linoff :

Use window functions:

select t.*,
       (case when max(num) over (partition by id) > 1 then 'A' else 'B' end) as label
from t;

If you actually want to update the table, aggregate and join:

update t
    set label = (case when max_num > 1 then 'A' else 'B' end)
    from (select id, max(num) as max_num
          from t 
          group by id
         ) tt
    where tt.id = t.id