如何根据一张表中的数据计算2个特定时期的收入差异？

Question

我使用的是SQL Server 2014，我有以下T-SQL查询，它应该计算基于2个特定时期的总收入的差异。执行计算的数据来自一个单一的SQL表。

概括地说，我在表中有一栏 T1 叫做 Revenue 和另一列称为 Month. 我需要找到2020年9月至2020年12月与2019年9月至2019年12月的收入差异。

我的T-SQL查询如下。

USE [MyDatabase]

;with cte1 as 
(
  SELECT 
    sum ([Revenue]) as 'Revenue Sep 19 - Dec 19'
  FROM
    [T1]
  WHERE
    [Month] between '2019-09-01' and '2019-12-01'
),
cte2 as (
  SELECT
    sum ([Revenue]) as 'Revenue Sep 20 - Dec 20'
  FROM
    [T1]
  WHERE
    [Month] between '2020-09-01' and '2020-12-01'
),
cte3 as (
  SELECT 
    cte2.[Revenue Sep 20 - Dec 20] as 'Total Revenue',
    'Sep 20 - Dec 20' as 'Period',
    '1' as 'ID'
  FROM
    [cte2]

  UNION ALL 

  SELECT 
    cte1.[Revenue Sep 19 - Dec 19] as 'Total Revenue',
    'Sep 19 - Dec 19' as 'Period',
    '1' as 'ID'
  FROM
    [cte1]
)
select a.[Total Revenue] - b.[Total Revenue]
from
  (select cte3.[Total Revenue] from [cte3] where cte3.[Period] = 'Sep 20 - Dec 20') a
  JOIN
  (select cte3.[Total Revenue] from [cte3] where cte3.[Period] = 'Sep 19 - Dec 19') b
  ON b.[ID] = a.[ID]

我的查询基于以下内容： 如何计算同一张表的不同行数组之间的差异

然而，当运行我的查询时，我得到以下错误信息。

Invalid column name 'ID'.

我不知道我在这里做错了什么。列名是不是 ID 在于 cte3?

Answer 1

ID 必须存在于选择列表中的两个 a 和 b 以使它能被人看到 join:

from
  (select cte3.ID, cte3.[Total Revenue] from [cte3] where cte3.[Period] = 'Sep 20 - Dec 20') a
  JOIN
  (select cte3.ID, cte3.[Total Revenue] from [cte3] where cte3.[Period] = 'Sep 19 - Dec 19') b
  ON b.[ID] = a.[ID]

Answer 2

你的查询比必要的复杂得多。

SELECT (SUM(CASE WHEN [Month] between '2020-09-01' and '2020-12-01' THEN [Revenue] ELSE 0 END) -
        SUM(CASE WHEN [Month] between '2019-09-01' and '2019-12-01' THEN [Revenue] ELSE 0 END)
       ) as difference
FROM [T1];

你的版本有三个CTE，一个 UNION ALL，两个子查询，以及一个 JOIN.

注意：不要用单引号来划分列别名。

• 不要用单引号来分隔列别名。 只对字符串和日期使用单引号。
• 命名列别名时，不需要对它们进行转义。 例如，使用下划线代替空格。
• 不要使用无意义的表别名，如 a 和 b. 使用有意义的，如 t2019.

Answer 3

当使用 SUB QUERY 你必须提到你所提到的列在 加入 条件也

SELECT A.[TOTAL REVENUE] - B.[TOTAL REVENUE]
FROM
(SELECT CTE3.[TOTAL REVENUE],CTE3.ID FROM [CTE3] WHERE CTE3.[PERIOD] = 'SEP 20 - DEC 20') A
JOIN
(SELECT CTE3.[TOTAL REVENUE],CTE3.ID FROM [CTE3] WHERE CTE3.[PERIOD] = 'SEP 19 - DEC 19') B ON B.[ID] = A.[ID]

Answer 4

试试下面这个代码，它将帮助你

USE [MyDatabase]

select sum([Revenue Sep 20 - Dec 20])-sum([Revenue Sep 19 - Dec 19]) as revenue_diffrence 
from
(
SELECT 
    [Revenue] as 'Revenue Sep 19 - Dec 19', 0 as 'Revenue Sep 20 - Dec 20'
  FROM
    [T1]
  WHERE
    [Month] between '2019-09-01' and '2019-12-01'

union all
SELECT
    0 as 'Revenue Sep 19 - Dec 19',[Revenue] as 'Revenue Sep 20 - Dec 20'
  FROM
    [T1]
  WHERE
    [Month] between '2020-09-01' and '2020-12-01'
)a