sql - SQL SERVER 中使用 PIVOT 命令的行到列（将 NULL 值替换为 0 并显示所有值的总和）

Question

我想在 Sql Server 中显示行到列。我已经看到了其他问题，但这些列在数据透视表中是硬编码的，但我的列将是动态的。到目前为止我所取得的成就。如屏幕截图所示，我能够将行转换为列，但我无法完成的几件事..需要你的帮助

1. 将所有列中的 NULL 替换为 0
2. 需要再添加 1 列，它将显示除 companyID 之外的所有列的总和

我的 SQL 代码：

DECLARE @Columns VARCHAR(MAX)
DECLARE @Convert VARCHAR(MAX)

SELECT  @Columns = STUFF((
           SELECT '],[' + ErrClassfn
             from ArchimedesTables.dbo.PM_ErrClassificationSetup
            WHERE CONVERT(VARCHAR(10), ISNULL(EndDate, GETDATE()), 101) 
                          Between CONVERT(VARCHAR(10), GETDATE(), 101)
                          AND     CONVERT(VARCHAR(10), GETDATE(), 101)
            ORDER BY '],[' + CONVERT(VARCHAR(MAX), ID) ASC
              FOR
                  XML PATH('')
           ), 1, 2, '') + ']'

SET @Convert = 'SELECT  * INTO #mynewTable FROM
(
    SELECT COUNT(WQ.ErrClassfnID) as ErrorCount, UPPER(WQ.CompanyID) as CompanyID, 
           PME.ErrClassfn as ErrorName
      FROM Version25.dbo.WF_Quality AS WQ
      LEFT JOIN ArchimedesTables.dbo.PM_ErrClassificationSetup as PME
           ON WQ.ErrClassfnID = PME.ID
     GROUP BY
          UPPER(CompanyID), ErrClassfn
) Quality PIVOT ( SUM(ErrorCount)  For ErrorName IN (' + @Columns
    + ')) as PivotTable  SeLeCt * FROM #mynewTable'

EXEC(@Convert)

Answer 1

您可以更改动态枢轴的列名称等，类似于：

DECLARE @ColumnsNull VARCHAR(MAX)
DECLARE @Columns VARCHAR(MAX)
DECLARE @Convert VARCHAR(MAX)

SELECT  @ColumnsNull = STUFF((SELECT ', IsNull(' + QUOTENAME(ErrClassfn) +', 0) as ['+ rtrim(ErrClassfn)+']' 
                   from ArchimedesTables.dbo.PM_ErrClassificationSetup
                   WHERE CONVERT(VARCHAR(10), ISNULL(EndDate, GETDATE()), 101) 
                                  Between CONVERT(VARCHAR(10), GETDATE(), 101)
                                  AND     CONVERT(VARCHAR(10), GETDATE(), 101)
                   ORDER BY ID ASC
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'')

SELECT  @Columns = STUFF((
           SELECT '],[' + ErrClassfn
             from ArchimedesTables.dbo.PM_ErrClassificationSetup
            WHERE CONVERT(VARCHAR(10), ISNULL(EndDate, GETDATE()), 101) 
                          Between CONVERT(VARCHAR(10), GETDATE(), 101)
                          AND     CONVERT(VARCHAR(10), GETDATE(), 101)
            ORDER BY '],[' + CONVERT(VARCHAR(MAX), ID) ASC
              FOR
                  XML PATH('')
           ), 1, 2, '') + ']'


SET @Convert = 'SELECT  CompanyID, '+ @ColumnsNull + '
                INTO #mynewTable 
                FROM
                (
                    SELECT COUNT(WQ.ErrClassfnID) as ErrorCount, UPPER(WQ.CompanyID) as CompanyID, 
                           PME.ErrClassfn as ErrorName
                      FROM Version25.dbo.WF_Quality AS WQ
                      LEFT JOIN ArchimedesTables.dbo.PM_ErrClassificationSetup as PME
                           ON WQ.ErrClassfnID = PME.ID
                     GROUP BY
                          UPPER(CompanyID), ErrClassfn
                ) Quality PIVOT ( SUM(ErrorCount)  For ErrorName IN (' + @Columns
                    + ')) as PivotTable  SeLeCt * FROM #mynewTable'

EXEC(@Convert)

我建议先编写查询并让列正常工作，然后将数据添加到#temp 表中。这样调试会更容易。

您也可以SUM()以相同的方式创建一个字段，在其中动态构建它，然后将其添加到 final 中SELECT：

所以它可能是这样的，你可以添加到 final 中SELECT：

SELECT  @ColumnsTotal = STUFF((SELECT '+' + QUOTENAME(ErrClassfn) 
                   from ArchimedesTables.dbo.PM_ErrClassificationSetup
                   WHERE CONVERT(VARCHAR(10), ISNULL(EndDate, GETDATE()), 101) 
                                  Between CONVERT(VARCHAR(10), GETDATE(), 101)
                                  AND     CONVERT(VARCHAR(10), GETDATE(), 101)
                   ORDER BY ID ASC
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'')