分组找ID

select t.name1,t.name2,t.zx
from (select row_number()over(partition by name1 order by zx desc)rn, test.* from test)t
where t.rn=1;