在 SQL Server 中如何透视多个列

Posted 2023-05-08

【中文标题】在 SQL Server 中如何透视多个列

【英文标题】：In SQL Server how to Pivot for multiple columns

【发布时间】：2016-06-28 05:09:46

【问题描述】：

这是我的示例表，我想旋转类别列并将销售额、库存和目标作为行来获取

我想要这种形式的示例输出，如下所示，其中类别代替列，列代替行

【问题讨论】：

@saad 您需要为您发布的问题做出回应。否则以后没人会帮你了

【参考方案1】：

您必须更改下一个 Pivot 语句的列名称。

喜欢

SELECT
*
FROM
(
  SELECT 
   Branch,
   Category,
   Category+'1' As Category1,
   Category+'2' As Category2,
   Sales, 
   Stock, 
   Target
  FROM TblPivot
 ) AS P

 -- For Sales
 PIVOT
 (
   SUM(Sales) FOR Category IN ([Panel], [AC], [Ref])
 ) AS pv1

 -- For Stock
 PIVOT
 (
   SUM(Stock) FOR Category1 IN ([Panel1], [AC1], [Ref1])
 ) AS pv2

 -- For Target
 PIVOT
 (
   SUM(Target) FOR Category2 IN ([Panel2], [AC2], [Ref2])
 ) AS pv3
 GO

---

你现在准备好了....

您可以使用 pv3 的聚合按您需要的列进行汇总和分组。

【讨论】：

这里的关键点是通过在末尾附加1 或2 来创建新的类别值。如果不这样做，透视查询将无法正常工作。

这一项应该标记为正确答案。

【参考方案2】：

样本表：

DECLARE @Table1 TABLE 
    (Branch varchar(9), Category varchar(9), Sales INT,Stock INT,Target INT)
;

INSERT INTO @Table1
    (Branch, Category, Sales, Stock,Target)
VALUES
    ( 'mumbai', 'panel', 10,4,15),
    ( 'mumbai', 'AC', 11,7,14),
    ( 'mumbai', 'Ref', 7,2,10),
    ( 'Delhi', 'panel',20,4,17),
    ( 'Delhi', 'AC', 5,2,12),
    ( 'Delhi', 'Ref', 10,12,22)
;

在 SQL SERVER 脚本中：

  Select BRANCH,COL,[panel],[AC],[Ref] from (
    select Branch,Category,COL,VAL from @Table1
    CROSS APPLY (VALUES ('Sales',Sales),
    ('Stock',Stock),
    ('Target',Target))CS (COL,VAL))T
    PIVOT (MAX(VAL) FOR Category IN ([panel],[AC],[Ref]))PVT
ORDER BY Branch DESC

【参考方案3】：

以下应该可以工作，

select * FROM
(
  SELECT 
   Branch,
   Category,
   Sales, 
   Stock, 
   Target
  FROM Table1
 ) AS P
 unpivot
 (
 [Value] FOR [OutPut] IN (sales,stock,[target])
 )unpvt
 pivot
 (
 max([Value]) for  Category in (Panel,AC,Ref) 
 )pvt
order by Branch Desc

【参考方案4】：

试试下面的解决方案

  -- Applying pivoting on multiple columns
SELECT
*
FROM
(
  SELECT 
   Category,
   Sales, 
  FROM TblPivot
 ) AS P

 -- For Sales
 PIVOT
 (
   SUM(Sales) FOR Category IN ([Panel], [AC], [Ref])
 ) AS pv1

union all

 -- For Stock
 SELECT
*
FROM
(
  SELECT 
   Category,
   Stock, 
  FROM TblPivot
 ) AS P

 PIVOT
 (
   SUM(Stock) FOR Category IN ([Panel], [AC], [Ref])
 ) AS pv2

union all

 -- For Target
 SELECT
*
FROM
(
  SELECT 
   Category,
   Target, 
  FROM TblPivot
 ) AS P

 PIVOT
 (
   SUM(Target) FOR Category IN ([Panel], [AC], [Ref])
 ) AS pv3
 GO

【讨论】：

我得到'无效的列名'类别'。可以请教吗？

【参考方案5】：

来自Answer1 & Answer2

（这不是 OP 的要求）

DECLARE @Table1 TABLE(Branch varchar(9), Category varchar(9), Sales INT,Stock INT,Target INT);

INSERT INTO @Table1
    (Branch, Category, Sales, Stock,Target)
VALUES
    ( 'mumbai', 'panel', 10,4,15),
    ( 'mumbai', 'AC', 11,7,14),
    ( 'mumbai', 'Ref', 7,2,10),
    ( 'Delhi', 'panel',20,4,17),
    ( 'Delhi', 'AC', 5,2,12),
    ( 'Delhi', 'Ref', 10,12,22);

SELECT
    Branch,
    SUM(Panel) As PanelSales,SUM([AC]) As ACSales,SUM([Ref]) As RefSales,
    SUM(Panel1) As PanelStock,SUM([AC1]) As ACStock,SUM([Ref1]) As RefStock,
    SUM(Panel2) As PanelTarget,SUM([AC2]) As ACTarget,SUM([Ref2]) As RefTarget
FROM
(
  SELECT 
   Branch,
   Category,
   Category+'1' As Category1,
   Category+'2' As Category2,
   Sales, 
   Stock, 
   Target
  FROM @Table1
 ) AS P

 -- For Sales
 PIVOT
 (
   SUM(Sales) FOR Category IN ([Panel], [AC], [Ref])
 ) AS pv1

 -- For Stock
 PIVOT
 (
   SUM(Stock) FOR Category1 IN ([Panel1], [AC1], [Ref1])
 ) AS pv2

 -- For Target
 PIVOT
 (
   SUM(Target) FOR Category2 IN ([Panel2], [AC2], [Ref2])
 ) AS pv3
 Group BY Branch
 GO

【参考方案6】：

您可以先UNPIVOT 数据使其成为平面表格，然后PIVOT 将类别转换为列：

SELECT *
FROM SampleTable
UNPIVOT (Amount FOR Output IN (Sales, Stock, Target)) upvt
PIVOT (SUM(Amount) FOR Category IN (Panel, AC, Ref)) pvt
ORDER BY Branch, Output;