在SQL Server 2014中将(文本)行转换为列

我正在尝试捕获特定组的行，然后将它们转换为命名脚本标头的列。

这是一个数据示例：

FORMULAS  | INSTRUCTIONS
----------+----------------------------------------------------
AF100120  | **ACETALDEHYDE WARNING - USE APPROPRIATE CAUTION**
AF100120  | ADD REMAINING ITEMS   
AF100120  | DISSOLVE ITEMS 1-2 IN 3 
AF100120  | IF HEAT USED, COOL TO ROOM TEMPERATURE 
AF100120  | MIX UNTIL HOMOGENEOUS                                                                               
AF100120  | MIXING TIMES VARY WITH BATCH SIZE                                                                   
AF100997  | 1) DISSOLVE THYMOL CRYSTALS IN ETHYL ALCOHOL. MIX WELL.
AF100997  | 2) ADD REMAINING ITEMS AND MIX UNTIL UNIFORM.
AF100997  | 3) FILTER IF NOT CLEAR

此脚本的所需输出为：

FORMULAS | INSTRUCTION #1 |INSTRUCTION #2 | INSTRUCTION #3 |INSTRUCTION #4 |INSTRUCTION #5|INSTRUCTION #6

AF100120 |**ACETALDEHYDE..| ADD REMAINING.| DISSOLVE ITEM. |IF HEAT USED.. |MIX UNTIL H   |MIXING TIM...
AF100997 |1) DISSOLVE THYM| 2) ADD REMAINI| FILTER IF NOT  | NULL          |NULL          |NULL          

如果有人能帮我编写代码，我将不胜感激：

SELECT *
FROM
    (SELECT 
         A.FormulaId AS FORMULAS, 
         CAST(B.Instruction AS NVARCHAR(100)) AS [INSTRUCTIONS]
     FROM 
         Formulas AS A
     LEFT JOIN 
         FormulaProcedures AS B ON A.FormulaId = B.FormulaId
     WHERE 
         B.ProcedureType = 3 
         AND PhaseType = 2 
         AND YEAR(A.VersionDate) >= '2018' 
         AND A.PrimaryVersion = 1 
         AND A.FormulaId IN('AF100120', 'AF100997')
     GROUP BY  
         A.FormulaId, CAST(B.Instruction AS NVARCHAR(100))
    ) AS SourceTable 
PIVOT
    (MAX([INSTRUCTIONS]) 
         FOR [INSTRUCTIONS] IN ([INSTRUCTION #1], [INSTRUCTION #2],
                                [INSTRUCTION #3], [INSTRUCTION #4],
                                [INSTRUCTION #5], [INSTRUCTION #6])) AS PivotTable;

谢谢！

---

解决方案

可以使用动态透视包含ROW_NUMBER()函数，其值按formulas按instructions列分组，以便在迭代分析函数时对列名进行编号：

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

SELECT @cols = 
       STUFF((SELECT DISTINCT ',' + QUOTENAME(CONCAT('formulas',[rn])) AS formulas
                FROM 
                (
                 SELECT f.*, 
                        ROW_NUMBER() OVER 
                                 (PARTITION BY [formulas] ORDER BY [instructions]) AS rn
                   FROM [Formulas_Tab] f
                ) ff
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'');

SET  @query = 
 N'SELECT *  
     FROM
     (
       SELECT f.*, 
              CONCAT(''formulas'',
                     ROW_NUMBER() OVER 
                              (PARTITION BY [formulas] ORDER BY [instructions])) AS rn
         FROM [Formulas_Tab] f
      ) ff
    PIVOT 
    (
     MAX([instructions]) FOR [rn] IN (' + @cols + N')
    ) p '

EXEC sp_executesql @query;

Demo

如果您安装的数据库的版本是2017+，那么您将使用STRING_AGG()函数来确定@cols变量为

SELECT @cols =
SELECT STRING_AGG(QUOTENAME(CONCAT('formulas', [ rn ])), ',') AS formulas
  FROM 
  (
   SELECT DISTINCT ROW_NUMBER() OVER
          ( PARTITION BY [ formulas ] ORDER BY [ instructions ]) AS rn
     FROM [ Formulas_Tab ] f
  ) ff