我需要一个包含前几行总和的新行答案

【问题标题】：I need a new line with the sum of the previous lines我需要一个包含前几行总和的新行
【发布时间】：2018-04-27 09:47:38
【问题描述】：

我想要所有表格行，并且每当银行字段中的值更改为带有字段小计的新行时：value1、value2、value3

我想使用 SQL 查询执行以下操作：

Create Table #MyTable (
    bank nvarchar(255),
    account nvarchar(255),
    value1 decimal(18,2),
    value2 decimal(18,2),
    value3 decimal(18,2),
);

INSERT INTO #MyTable
  (bank, account, value1, value2, value3)
VALUES
  ('VR-Bank', 'Kto-A', 1, 2, 3),
  ('VR-Bank', 'Kto-B', 3, 2, 1),
  ('VR-Bank', 'Kto-C', 4, 4, 5),
  ('NOSPA', 'Kto-X', 1, 1, 1),
  ('NOSPA', 'Kto-Y', 4, 3, 2);

我想要没有先前总和的总和，因为这些值已经存在

Select 
    bank,
    account,
    SUM(value1), -- I would like this without SUM()
    SUM(value2),
    SUM(value3)
from #MyTable
GROUP BY 
    GROUPING SETS ((bank), (account)) -- Without summing up this does not work

-- NULL Kto-A   1.00    2.00    3.00
-- NULL Kto-B   3.00    2.00    1.00
-- NULL Kto-C   4.00    4.00    5.00
-- NULL Kto-X   1.00    1.00    1.00
-- NULL Kto-Y   4.00    3.00    2.00
-- NOSPA    NULL    5.00    4.00    3.00
-- VR-Bank  NULL    8.00    8.00    9.00

我想实现以下目标：

-- ======================================

-- VR-Bank  Kto-A   1.00    2.00    3.00 -- here also the value for the bank
-- VR-Bank  Kto-B   3.00    2.00    1.00
-- VR-Bank  Kto-C   4.00    4.00    5.00
-- VR-Bank  NULL    8.00    8.00    9.00 -- here the SUM
-- NOSPA    Kto-X   1.00    1.00    1.00
-- NOSPA    Kto-Y   4.00    3.00    2.00
-- NOSPA    NULL    5.00    4.00    3.00 -- SUM for the next bank

Drop Table #MyTable;

我希望有人可以帮助我。

【问题讨论】：

SQL SERVER T-SQL Calculate SubTotal and Total by group的可能重复
GROUP BY GROUPING SETS ((bank, account), (bank))?
非常感谢，很合适。我之前尝试过 GROUP BY、ROLLUP、CUBE 等。当然，我还必须计算总和，这样 GROUPING SETS 才能正常工作。问题已为我解答。

标签： sql sql-server

【解决方案1】：

试试这个：

Select 
    bank,
    account,
    SUM(value1) value1, 
    SUM(value2) value2,
    SUM(value3) value3
from #MyTable
GROUP BY 
    GROUPING SETS ((bank, account, value1, value2, value3), (bank)) --

应该相当于：

select *
from (
    select * from #MyTable
    union all
    Select 
        bank,
        NULL account,
        SUM(value1) value1, 
        SUM(value2) value2,
        SUM(value3) value3
    from #MyTable
    GROUP BY bank
) x
order by 1, 2 desc

但第一个性能更高

【讨论】：

【解决方案2】：

如果您想要原始行并使用 grouping sets 将它们组合起来，请首先向表中添加唯一 id：

Create Table #MyTable (
    mytableId int identity(1, 1) primary key,
    bank nvarchar(255),
    account nvarchar(255),
    value1 decimal(18,2),
    value2 decimal(18,2),
    value3 decimal(18,2),
);

根据测量值来确保行是唯一的，我会感到不舒服。

那么你可以简单地做：

select bank, id, max(account) as account,
       sum(value1) as value1, sum(value2) as value2, sum(value3) as value3
from #MyTable
group by grouping sets ( (bank, id), (bank) );

如果您的意图是 bank/account 在组合中是唯一的，那么您可以这样做：

SELECT bank,account,
       SUM(value1), -- I would like this without SUM()
       SUM(value2),
       SUM(value3)
FROM #MyTable
GROUP BY GROUPING SETS ((bank, account), (account));

【讨论】：

【解决方案3】：

SELECT (CASE WHEN GROUPING(bank) = 0 
             AND  GROUPING(account) = 1 
             THEN 'Total ' + bank
             ELSE bank
        END) AS bank,
        account,
        SUM(value1) AS value1,
        SUM(value2) AS value2,
        SUM(value3) AS value3
FROM MyTable 
GROUP BY GROUPING SETS((bank, account,value1, value2, value3),
                       (bank)
                      );

输出

bank            account value1  value2  value3
NOSPA           Kto-X   1       1       1
NOSPA           Kto-Y   4       3       2
Total NOSPA     (null)  5       4       3
VR-Bank         Kto-A   1       2       3
VR-Bank         Kto-B   3       2       1
VR-Bank         Kto-C   4       4       5
Total VR-Bank   (null)  8       8       9

演示

http://sqlfiddle.com/#!18/2c3b9/14

【讨论】：