文本值的 SQL 数据透视表答案

【问题标题】：SQL pivot table on text value文本值的 SQL 数据透视表
【发布时间】：2018-09-03 09:57:30
【问题描述】：

我正在尝试在 SQL 中转换以下数据集。

SELECT  ParticipantID,
        ContactReasonTypeID,
        Completed = CASE WHEN Returned IS NOT NULL
                        THEN    'Completed'
                        ELSE 'Not completed'
                    END
FROM    ContactReason
WHERE   ContactReasonTypeID = 40 -- M4
UNION
SELECT  ParticipantID,
        ContactReasonTypeID,
        Completed = CASE WHEN Returned IS NOT NULL
                        THEN    'Completed'
                        ELSE 'Not completed'
                    END
FROM    ContactReason
WHERE   ContactReasonTypeID = 80 -- M8

结果如下：

ParticipantID | QuestionnaireID  | Completed
1               40                 'Completed'
1               80                 'Not completed'
2               40                 'Completed'
2               80                 'Completed'
3               40                 'Completed'
3               80                 'Not completed'
4               40                 'Completed'
4               80                 'Not completed'

我想要的是像这样旋转结果（其中 M4 是 QuestionnaireID 40，M8 是 80）：

ParticipantID | M4           | M8
1               'Completed'    'Not completed'
2               'Completed'    'Completed'
3               'Completed'    'Not completed'
4               'Completed'    'Not completed'

我正在为如何旋转表格而苦苦挣扎。我当前的 SQL 如下并导致错误。我尝试根据其他帖子切换列，但无法确定需要什么：

SELECT  *
FROM    (
    SELECT  ParticipantID,
            ContactReasonTypeID,
            Completed = CASE WHEN Returned IS NOT NULL
                            THEN    'Completed'
                            ELSE 'Not completed'
                        END
    FROM    ContactReason
    WHERE   ContactReasonTypeID = 40
    UNION
    SELECT  ParticipantID,
            ContactReasonTypeID,
            Completed = CASE WHEN Returned IS NOT NULL
                            THEN    'Completed'
                            ELSE 'Not completed'
                        END
    FROM    ContactReason
    WHERE   ContactReasonTypeID = 80
) AS tbl
PIVOT (
    MAX(Returned)
    FOR ContactReasonTypeID IN ([M4],[M8])
) AS pvt
ORDER BY ParticipantID

【问题讨论】：

标签： sql sql-server tsql pivot

【解决方案1】：

您可以为此使用命令PIVOT

DECLARE @t table(ParticipantID INT, QuestionnaireID INT, Completed varchar(20))
INSERT @t values
(1,40,'Completed'),(1,80,'Not completed'),(2,40,'Completed'),
(2,80,'Completed'),(3,40,'Completed'),(3,80,'Not completed'),
(4,40,'Completed'),(4,80,'Not completed')


SELECT ParticipantID, [40] M4, [80] M8
FROM @t
PIVOT
  (min(COMPLETED)  
FOR QuestionnaireID
  in([40],[80])  
)AS p
ORDER BY ParticipantID

结果：

ParticipantID   M4          M8
1               Completed   Not completed
2               Completed   Completed
3               Completed   Not completed
4               Completed   Not completed

【讨论】：

【解决方案2】：

你可以试试下面的代码：

SELECT  DISTINCT CR.ParticipantID,
        (SELECT CASE
                    WHEN Returned IS NOT NULL
                    THEN 'Completed'
                    ELSE 'Not completed'
                END M4
           FROM ContactReason
          WHERE ContactReasonTypeID = 40 --M4
            AND ParticipantID = CR.ParticipantID) M4,
        (SELECT CASE
                    WHEN Returned IS NOT NULL
                    THEN 'Completed'
                    ELSE 'Not completed'
                END M4
           FROM ContactReason
          WHERE ContactReasonTypeID = 80 --M8
            AND ParticipantID = CR.ParticipantID) M8
FROM    ContactReason CR
WHERE   CR.ContactReasonTypeID IN (40, 80) -- M4 and M8

【讨论】：

这种方法效果很好，但是由于使用了数据透视表，我标记了选定的答案。

【解决方案3】：

你可以试试这个：

SELECT q.ParticipantID, 
                  CASE WHEN min(q.QuestionnaireID) = 40
                        THEN min(q.Completed)
                    END as M4,
                   CASE WHEN max(q.QuestionnaireID) = 80
                        THEN max(q.Completed)
                    END as M8
  FROM
(
SELECT  ParticipantID,
        ContactReasonTypeID,
        Completed = CASE WHEN Returned IS NOT NULL
                        THEN    'Completed'
                        ELSE 'Not completed'
                    END
FROM    ContactReason
WHERE   ContactReasonTypeID = 40 -- M4
UNION
SELECT  ParticipantID,
        ContactReasonTypeID,
        Completed = CASE WHEN Returned IS NOT NULL
                        THEN    'Completed'
                        ELSE 'Not completed'
                    END
FROM    ContactReason
WHERE   ContactReasonTypeID = 80
) q
GROUP BY ParticipantID;

在Demo 内部，将内部查询q 视为tab2。

【讨论】：