select a.id,a.job,a.score,a.t_rank from( select id,job,score ,(row_number() over(partition by job order by score desc))as t_rank ,floor((count() over(partition by job)+1)/2) as r_min ,ceiling((count() over(partition by job)+1)/2) as r_max from grade )a where a.t_rank between r_min and r_max order by...