将 Varchar 值旋转到列答案

【问题标题】：Pivoting Varchar Value to column将 Varchar 值旋转到列
【发布时间】：2021-03-24 15:01:48
【问题描述】：

我正在尝试在 SQL Server 2016 中转换此数据集

--数据集

CarModel    Tire              MakeDate   Owner

Subaru       MRF             20200101     John D
Subaru       Goodyear        20200101     John D
Subaru       MRF             20200101     John D
Subaru       MRF             20200101     John D

输出结果应该是：

CarModel Tire1 Tire2     Tire3 Tire4 MakeDate    Owner
Subaru   MRF   GoodYear  MRF    MRF   20200101    John D

我试过了，但没有成功。如果我犯了任何错误，请告诉我吗？

create table #temp_Dataset 
(Carmodel varchar(10), Tire varchar(10), MakeDate int, [Owner] Varchar(10))

insert into #temp_Dataset
(Carmodel, Tire, MakeDate, [Owner])
values 
('Subaru',       'MRF',             20200101 ,    'John D'),
('Subaru',       'Goodyear',        20200101 ,    'John D'),
('Subaru',       'MRF',             20200101 ,    'John D'),
('Subaru',       'MRF',             20200101 ,    'John D')

Select * from( 
 SELECT CarModel,
       Tire,
       MakeDate,
       [Owner]
FROM #temp_Dataset)a
pivot (max(carmodel) for Tire in (Tire1, Tire2, Tire3, Tire4))piv

drop table #temp_Dataset

【问题讨论】：

标签： sql sql-server pivot sql-server-2016

【解决方案1】：

您错过了枚举轮胎并在 MAX 聚合中使用了错误的参数 ([Carmodel])。更喜欢使用诸如

SELECT * 
  FROM( 
       SELECT *,
              CONCAT('Tire',ROW_NUMBER() OVER 
             (PARTITION BY [Carmodel], [MakeDate], [Owner] ORDER BY [CarModel])) AS [rn]
         FROM #temp_Dataset
      ) AS ds
  PIVOT(
        MAX([Tire]) FOR [rn] IN ([Tire1],[Tire2],[Tire3],[Tire4])
       ) AS piv

顺便说一句，在当前情况下，四个轮胎将任意排序，您可以通过更改ORDER BY 子句中的列来随意更改。

您甚至可以使旋转动态化，以便在按其他列分组时添加或减去轮胎的值，例如

DECLARE @cols  AS NVARCHAR(MAX),  @query AS NVARCHAR(MAX)
                
SELECT @cols = 
       STUFF((SELECT DISTINCT ',' + QUOTENAME(CONCAT('Tire',[rn])) AS formulas
                FROM 
                (
                 SELECT *, 
                        ROW_NUMBER() OVER 
                        (PARTITION BY [Carmodel], [MakeDate], [Owner] ORDER BY [CarModel]) AS rn
                   FROM [#temp_Dataset]
                ) ff
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'');
        
SET  @query =
          N'SELECT * 
              FROM( 
                   SELECT *,
                          CONCAT(''Tire'',
                          ROW_NUMBER() OVER (PARTITION BY [Carmodel], [MakeDate], [Owner] ORDER BY [CarModel])) AS [rn]
                     FROM [#temp_Dataset]
                  ) AS ds
             PIVOT(
                   MAX([Tire]) FOR [rn] IN (' + @cols + N' )
                  ) AS piv'

EXEC sp_executesql @query;

Demo

【讨论】：

【解决方案2】：

你可以使用条件聚合和row_number():

select carmodel, makedate, owner,
       max(case when seqnum = 1 then tire end) as tire_1,
       max(case when seqnum = 2 then tire end) as tire_2,
       max(case when seqnum = 3 then tire end) as tire_3,
       max(case when seqnum = 4 then tire end) as tire_4
from (select t.*,
             row_number() over (partition by carmodel, makedate, owner order by ?) as seqnum
      from #temp_Dataset t
     ) t
group by carmodel, makedate, owner;

? 用于列的排序。您希望结果的顺序并不明显——SQL 表表示无序集。

【讨论】：