Question

假设我有如下表格：

CustomerAdresses表：

Shop        Name        Surname     Adress
----------------------------------------------
SHOP1       name1       surname1    adress1_1 
SHOP2       name1       surname1    adress1_2 
SHOP1       name2       surname2    adress2_1 
SHOP3       name2       surname2    adress2_3 
SHOP3       name1       surname1    adress1_3

和

ShopPriority表：

Shop   Priority
---------------
SHOP1       1
SHOP2       2
SHOP3       3

我正在尝试选择具有最高优先级的客户地址（考虑到这是商店）

我写了sql：

select ca.shop, ca.name, ca.surname, ca.adress, sp.priority
from CustomerAdresses ca
join ShopPriority sp on sp.shop = ca.shop
group by ca.name, ca.surname, ca.adress, ca.shop, sp.priority

此查询向我返回如下结果：

Shop        name        surname     adress    priority
-------------------------------------------------------
SHOP1       name1       surname1    adress1_1   1
SHOP2       name1       surname1    adress1_2   2
SHOP3       name1       surname1    adress1_3   3
SHOP1       name2       surname2    adress2_1   1
SHOP3       name2       surname2    adress2_3   3

现在，我已经接近我所需要的。记录按名称和姓氏分组。现在，我只需要从每个组中获得最高优先级的记录。分组后我尝试订购记录，但是整个分组都崩溃了

Answer 1

您可以使用row_number()函数：

select t.*
from (select ca.shop, ca.name, ca.surname, ca.adress, sp.priority,
             row_number() over (partition by ca.shop order by sp.priority desc) as seq
      from CustomerAdresses ca join 
           ShopPriority sp 
           on sp.shop = ca.shop
      ) t
where seq = 1;

但是，top (1) with ties也很有用：

select top (1) with ties ca.shop, ca.name, ca.surname, ca.adress, sp.priority
from CustomerAdresses ca join 
     ShopPriority sp 
     on sp.shop = ca.shop
order by row_number() over (partition by ca.shop order by sp.priority desc);

Answer 2

使用ANSI标准的row_number()函数：

select spc.*
from (select ca.shop, ca.name, ca.surname, ca.adress, sp.priority,
             row_number() over (partition by ca.shop order by sp.priority desc) as seqnum
      from CustomerAdresses ca join
           ShopPriority sp
           on sp.shop = ca.shop
     ) spc
where seqnum = 1;

从每个组中选择具有最高价值的记录

2 个答案: