SQL：列和

让我们有以下示例表：

Person Quantity
A      1
B      2
C      3
D      4
E      5

结果应该是：

PersonAggregate
1     (0+Quantity of PersonA)=sumA
3     (sumA+Quantity of PersonB)=sumB
6     (sumB+Quantity of PersonC)=sumC
10    (sumC+Quantity of PersonD)=sumD
15    (sumD+Quantity of PersonE)

---

有可能在onq SQL查询中得到这个结果吗？

这显然会得到单独的和

select person, sum(quantity)
from sample
group by person
order by person

我不认为你想要的效果可以用固定的方式来实现。带有游标的过程语言，如T-SQL或PLSQL，可以很容易地实现这一点

---

我会编写一个存储过程并调用它。

这显然会得到单独的总和

select person, sum(quantity)
from sample
group by person
order by person

我不认为你想要的效果可以用固定的方式来实现。带有游标的过程语言，如T-SQL或PLSQL，可以很容易地实现这一点

---

我会编写一个存储过程并调用它。

大多数版本的SQL都支持将累积和作为窗口函数：

select person, sum(quantity) over (order by person) as cumesum
from sample;

也可以使用相关子查询执行此操作：

select s.person,
       (select sum(s2.quantity)
        from samples s2
        where s2.person <= s.person
       ) as cumesum
from sample s;

---

大多数SQL版本都支持将累计和作为窗口函数：

select person, sum(quantity) over (order by person) as cumesum
from sample;

也可以使用相关子查询执行此操作：

select s.person,
       (select sum(s2.quantity)
        from samples s2
        where s2.person <= s.person
       ) as cumesum
from sample s;

如果示例表中每人有多行，且有多个数量需要求和，则可以使用：

select curr.person, curr.sum_person + case when prev.person <> curr.person
                                           then prev.sum_person
                                           else 0 end as person_sum
  from (select person, sum(quantity) as sum_person
          from sample
         group by person) curr
 cross join (select person, sum(quantity) as sum_person
               from sample
              group by person) prev
 where prev.person =
       (select max(x.person) from sample x where x.person < curr.person)
    or curr.person = (select min(person) from sample)
 group by curr.person

如果示例表中每人只有一行，则可以更简单地使用：

select curr.person, curr.quantity +   case when prev.person <> curr.person
                                           then prev.quantity
                                           else 0 end as person_sum
  from sample curr
 cross join sample prev
 where prev.person =
       (select max(x.person) from sample x where x.person < curr.person)
    or curr.person = (select min(person) from sample)
 group by curr.person

小提琴：

如果样本每人有2+行：

select person,
       sum(quantity) over(order by person rows between 1 preceding and current row) as your_sum
  from sample
 order by person

select person,
       sum(sum_person) over(order by person rows between 1 preceding and current row) as your_sum
  from (select person, sum(quantity) as sum_person
          from sample
         group by person) x
 order by person

Fiddle:

如果样本表中每人有多行，且有多个数量需要求和，则可以使用：

select curr.person, curr.sum_person + case when prev.person <> curr.person
                                           then prev.sum_person
                                           else 0 end as person_sum
  from (select person, sum(quantity) as sum_person
          from sample
         group by person) curr
 cross join (select person, sum(quantity) as sum_person
               from sample
              group by person) prev
 where prev.person =
       (select max(x.person) from sample x where x.person < curr.person)
    or curr.person = (select min(person) from sample)
 group by curr.person

如果示例表中每人只有一行，则可以更简单地使用：

select curr.person, curr.quantity +   case when prev.person <> curr.person
                                           then prev.quantity
                                           else 0 end as person_sum
  from sample curr
 cross join sample prev
 where prev.person =
       (select max(x.person) from sample x where x.person < curr.person)
    or curr.person = (select min(person) from sample)
 group by curr.person

小提琴：

如果样本每人有2+行：

select person,
       sum(quantity) over(order by person rows between 1 preceding and current row) as your_sum
  from sample
 order by person

select person,
       sum(sum_person) over(order by person rows between 1 preceding and current row) as your_sum
  from (select person, sum(quantity) as sum_person
          from sample
         group by person) x
 order by person

---

Fiddle:

一个人是否有2+行的多个数量需要求和？我这样问是因为，如果不是，您的查询实际上不会涉及求和函数。相反，您只是水平添加两个值。您使用的是哪种DBMS？博士后？Oracle？一个人是否有2行以上的多个数量需要求和？我这样问是因为，如果不是，您的查询实际上不会涉及求和函数。相反，您只是水平添加两个值。您使用的是哪种DBMS？博士后？神谕