Question

问题：我需要为供应商找到顶级客户。顶级客户在所有订单上都有最大项目总和。如果供应商有2个顶级客户，那么顶级客户将使用最小订单ID，这意味着 - 谁首先创建订单并获得最大订单 - 它将是最高的。我正在写这个sql代码：

select s.s_id, c.c_id, min(o.o_id), count(*)
from suppliers s, clients c, orders o, items i
where s.s_id=c.id_s and c.c_id=o.id_c and o.o_id=i.id_o
group by s.s_id, c.c_id
order by 4 desc, 3

并获得此结果：http://imageshack.us/photo/my-images/148/32969388.jpg/

但我需要：http://imageshack.us/photo/my-images/842/51815927.jpg/

Answer 1

您需要对订单计数和ID进行排名。因此，您应该使用分析和内联视图。类似的东西：

select s_id
       , c_id
       , min_order_id
       , no_of_orders
from (
    select s_id
           , c_id
           , min_order_id
           , no_of_orders
           , rank() over (partition by s_id 
                          order by no_of_orders DESC, min_order_id ASC) rnk
    from (
        select s.s_id
               , c.c_id
               , min(o.o_id) as min_order_id
               , count(*) as no_of_orders
        from suppliers s, clients c, orders o, items i
        where s.s_id=c.id_s and c.c_id=o.id_c and o.o_id=i.id_o
        group by s.s_id, c.c_id
        )
    )    
where rnk=1
/

如何在Oracle中的一列中选择多个没有重复的列

1 个答案: