Question

我在SQL表中有几列，我想计算表中每一行的结果。我的数据看起来像这样：

data
name     value   adjustor1    adjustor2
Comp1    20       0.05        0.08
Comp2    80      -0.07        0.065

每行调整值的公式为：

adjusted_value = value*(1 + adjustor1)*(1 + adjustor2)*(100/sum(value))

所以调整后的输出应为：

data
name     adjusted_value
Comp1    22.25
Comp2    77.75

原始值总和为100，调整后的值也应该总和为100.我尝试了以下内容：

SELECT adjusted*(100/sum(adjusted)) 
FROM (
      SELECT value*(1+adjustor1)*(1+adjustor2) as adjusted 
      FROM data
) as result

给了我错误：ERROR: column "result.adjusted" must appear in the GROUP BY clause or be used in an aggregate function

虽然我只是这样做：

SELECT sum(adjusted) 
FROM (
       SELECT value*(1+adjustor1)*(1+adjustor2) as adjusted 
       FROM data
) as result

OR

SELECT adjusted
FROM (
      SELECT value*(1+adjustor1)*(1+adjustor2) as adjusted 
      FROM data
) as result

我可以得到总和或调整后的值，但不能同时得到两者。

Answer 1

列名value不是一个好选择，它是一个保留字。我将其更改为val。

这是您需要的查询：

WITH data(name,val,adjustor1,adjustor2) AS (
    VALUES
    ('Comp1'::text,20,0.05,0.08),
    ('Comp2',80,-0.07,0.065)
)
SELECT name,val,adjustor1,adjustor2,
       CAST(val*(1+adjustor1)*(1+adjustor2)*(100/sum(
         val*(1+adjustor1)*(1+adjustor2)
       ) OVER ()) AS numeric(10,3)) adjusted_value
  FROM data;

我必须承认，它有点笨拙，子查询使其更容易理解（SQL Fiddle）：

WITH data(name,val,adjustor1,adjustor2) AS (
    VALUES
    ('Comp1'::text,20,0.05,0.08),
    ('Comp2',80,-0.07,0.065)
)
SELECT name, CAST(adj*(100/sum(adj) OVER ()) AS numeric(8,3)) adjusted_value
  FROM (
    SELECT name,val,adjustor1,adjustor2,
           val*(1+adjustor1)*(1+adjustor2)  adj
      FROM data) s;

一些注意事项：

原始sum(adj)已替换为sum(adj) OVER ()，window functions的语法。
我还添加了[CAST()][3]以汇总值。

Answer 2

你可以在子查询中获得SUM然后你可以CROSS JOIN来实现你想要的东西

SELECT name, 
       value*(1+adjustor1)*(1+adjustor2)*(100/T.adjSum) as adjusted_value
FROM data
CROSS JOIN 
(SELECT SUM(value*(1+adjustor1)*(1+adjustor2)) as adjSum 
 FROM data) T

SQL查询：获取每行的调整值

2 个答案: