在 SQL 中使用“Pivot”将行转换为列

Posted 2023-04-13

【中文标题】在 SQL 中使用“Pivot”将行转换为列

【英文标题】：Convert Rows to columns using 'Pivot' in SQL

【发布时间】：2020-12-01 23:26:06

【问题描述】：

我已经阅读了 MS 数据透视表上的内容，但我仍然无法正确解决这个问题。

我对这些结果有看法

Invoice Date        Basecard    Item      Quantity  Subtotal    Taxes   Total
713938  09/11/2020  C90001      Desktop   14            2800       448      3248
713938  09/11/2020  C90001      Laptop    18.5      29091.25    4654.6  33745.85

我希望它以数据透视表的形式出现，如下所示：

Invoice Date        Basecard    Laptop  Desktop   Subtotal  Taxes   Total
713938  09/11/2020  C90001      18.5    14        31891.25  5102.6  36993.85

这是我的查询

SELECT * FROM (
    SELECT Invoice, Date, BaseCard, Item, sum(Quantity) Qty, sum(Subtotal) Subtotal
    FROM MyView
    Group by Invoice, Date, BaseCard, Item, Subtotal
    
) Resultados
PIVOT (
        Sum(Qty)
        FOR Item
        IN (
            [Laptop], [Desktop]         
        )
) AS PivorTable
Group by Invoice, Date, BaseCard, Laptop, Desktop, Subtotal

结果

Invoice Date      baseCard  subtotal    Laptop  Desktop
713938  2020-11-09 C90001   2800.00     NULL    14.000000
713938  2020-11-09 C90001   7076.25     4.50    NULL
713938  2020-11-09 C90001   22015.000   14.00   NULL

【参考方案1】：

假设你有这样一张桌子

CREATE TABLE [InvoiceInfo]( 
                           [Invoice]  INT, 
                           [Date]     DATE, 
                           [Basecard] VARCHAR(100), 
                           [Item]     VARCHAR(100),
                           [Quantity] FLOAT,
                           [Price]    FLOAT
                          )

您从中创建视图。考虑到这一点，我可以建议您使用条件聚合来简化您的数据透视，并根据当前存在的[Item] 值转换为动态格式以使其灵活，例如

DECLARE @cols  AS NVARCHAR(MAX),  @query AS NVARCHAR(MAX)

SET @cols = ( SELECT STRING_AGG(
                     CONCAT('SUM(CASE WHEN [Item]=''',[Item],
                            ''' THEN [Quantity] END) AS [',[Item],']'),',') 
                FROM (SELECT DISTINCT [Item]
                        FROM [InvoiceInfo] ) i );

SET @query = CONCAT(
    N'SELECT [Invoice], [Date], [BaseCard],', @cols , 
    N'      ,SUM([Price]*[Quantity]) AS [Subtotal],
             SUM(ROUND([Price]*[Quantity]*.16, 2)) AS [Taxes],
             SUM([Price]*[Quantity])+SUM(ROUND([Price]*[Quantity]*.16, 2)) AS [Total]
        FROM [InvoiceInfo] i
       GROUP BY [Invoice], [Date], [BaseCard]');

EXEC sp_executesql @query; 

Invoice Date        BaseCard    Desktop Laptop  Subtotal    Taxes   Total
713938  2020-11-09  C90001      14      18.5    31891.25    5102.6  36993.85

更新：根据您最近声明的旧数据库版本，您可以使用以下代码，包括 FOR XML PATH 和 STUFF 来替代 STRING_AGG

DECLARE @cols AS NVARCHAR(MAX),  @query AS NVARCHAR(MAX);
                        
SELECT @cols = 
       STUFF((SELECT DISTINCT ',' + 
                     CONCAT('SUM(CASE WHEN [Item]=''',[Item],
                                            ''' THEN [Quantity] END) AS [',[Item],']') 
                  AS formulas
                FROM 
                (
                 SELECT DISTINCT [Item]
                   FROM [InvoiceInfo] f
                ) ff
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'');
       
SET @query = CONCAT(
    N'SELECT [Invoice], [Date], [BaseCard],', @cols , 
    N'      ,SUM([Price]*[Quantity]) AS [Subtotal],
             SUM(ROUND([Price]*[Quantity]*.16, 2)) AS [Taxes],
             SUM([Price]*[Quantity])+SUM(ROUND([Price]*[Quantity]*.16, 2)) AS [Total]
        FROM [InvoiceInfo] i
       GROUP BY [Invoice], [Date], [BaseCard]');

EXEC sp_executesql @query; 

Demo

【讨论】：

感谢您的帮助 Barbaros，但我有 SQL Server 2015 并且在运行查询时出现此错误。 -- STRING_AGG 不是可识别的内置函数名称 -- CONCAT 不是可识别的内置函数名称

嗨@Andros，不客气。由于您的数据库版本，我已更新。

对不起 Babaros，我的错误，我的 SQL 版本是 2005。我试图让它工作，但我做不到。你帮了大忙

【参考方案2】：

您也只需要总结这些关键数据。您的行已拆分，因为您已将 subtotal 列添加到分组依据。

db<>fiddle

with a as (
  select 713938 as invoice, convert(date, '2020-11-09', 23) as dt, 'C90001' as Basecard, 'Desktop' as item, 14 as quantity, 2800 as subtotal, 448 as taxes, 3248 as total
  union all
  select 713938, convert(date, '2020-11-09', 23), 'C90001', 'Laptop', 18.5, 29091.25, 4654.6, 33745.85
)
select
  invoice,
  dt,
  basecard,
  sum([Desktop]) as desctop,
  sum([Laptop]) as laptop,
  sum(subtotal) as subtotal,
  sum(taxes) as taxes,
  sum(total) as total
from a
pivot (
  sum(quantity) for item in ([Desktop], [Laptop])
) as q
group by
  invoice,
  dt,
  basecard

invoice | dt         | basecard | desctop | laptop | subtotal |  taxes |    total
------: | :--------- | :------- | ------: | -----: | -------: | -----: | -------:
 713938 | 2020-11-09 | C90001   |    14.0 |   18.5 | 31891.25 | 5102.6 | 36993.85

【讨论】：

感谢您的帮助，这几乎可以完美运行，如何转换为动态格式以使其灵活？因为我有更多的项目

@Andros 没有办法在静态 SQL 中使其动态化（我认为因为解析器在执行之前应该知道确切的列名）。因此，唯一的方法是使用动态 SQL，正如另一个答案中所述。还有很多例子：one、two 和，糟糕但有效，从第二个得到，three。