具有动态标头的SQL PIVOT

我有一个SQL数据库表，它的结构如下，我想让我的用户可以选择以垂直或水平方式导出数据

SELECT [MEA_ID], [MEA_DateTime], [PTR_ID], [EXP_ID], [DUN_ID], [MEA_Value]
FROM [MeasurementData];

|--------|----------------------|--------|--------|--------|-----------|
| MEA_ID |     MEA_DateTime     | PTR_ID | EXP_ID | DUN_ID | MEA_Value |
|--------|----------------------|--------|--------|--------|-----------|
| 1      | 2009-08-10 00:00:00  | 24     | 14     | 2      | 15.1      |
| 2      | 2009-08-10 00:00:00  | 24     | 14     | 3      | 14.3      |
| 3      | 2009-08-10 00:00:00  | 24     | 14     | 4      | 16.7      |
| 4      | 2009-08-10 00:00:10  | 24     | 15     | 2      | 13.0      |
| 5      | 2009-08-10 00:00:10  | 24     | 15     | 4      | 13.4      |
| 6      | 2009-08-10 00:00:20  | 24     | 16     | 2      | 17.8      |
| 7      | 2009-08-10 00:00:20  | 24     | 16     | 3      | 17.7      |
| 8      | 2009-08-10 00:00:20  | 24     | 16     | 4      | 16.2      |
| 9      | 2009-08-10 00:00:00  | 25     | 14     | 3      | 34.0      |
| 10     | 2009-08-10 00:00:00  | 25     | 14     | 4      | 19.0      |
| 11     | 2009-08-10 00:00:10  | 25     | 15     | 2      | 22.1      |
| 12     | 2009-08-10 00:00:10  | 25     | 15     | 3      | 23.1      |
| 13     | 2009-08-10 00:00:20  | 25     | 16     | 2      | 24.6      |
| 14     | 2009-08-10 00:00:20  | 25     | 16     | 3      | 18.3      |
| 15     | 2009-08-10 00:00:20  | 25     | 16     | 4      | 18.2      |

上表将是垂直输出

MEA_DateTime、PTR_ID、EXP_ID和DUN_ID的每个组合都是唯一的，因此对于给定的组合，始终只能有一行。我试图实现的是水平旋转DUN_ID，以便更好地比较值

应该是这样的：

SELECT [MEA_DateTime], [PTR_ID], [EXP_ID], [MEA_Value]
FROM [MeasurementData]
PIVOT
(
    SUM([MEA_Value])
    FOR [DUN_ID] IN (????)
);

                                          DUN_ID  DUN_ID  DUN_ID
                                             |       |       |
                                             v       v       v
|----------------------|--------|--------|-------|-------|-------|
|     MEA_DateTime     | PTR_ID | EXP_ID |   2   |   3   |   4   |
|----------------------|--------|--------|-------|-------|-------|
| 2009-08-10 00:00:00  | 24     | 14     | 15.1  | 14.3  | 16.7  |
| 2009-08-10 00:00:10  | 24     | 15     | 13.0  | NULL  | 13.4  |
| 2009-08-10 00:00:20  | 24     | 16     | 17.8  | 17.7  | 16.2  |
| 2009-08-10 00:00:00  | 25     | 14     | NULL  | 34.0  | 19.0  |
| 2009-08-10 00:00:10  | 25     | 15     | 22.1  | 23.1  | NULL  |
| 2009-08-10 00:00:20  | 25     | 16     | 24.6  | 18.3  | 18.2  |

我试着让它以一个支点工作，但不幸的是，以前从未做过类似的事情，也没有太多东西可以展示。根据我所了解的情况，您需要事先知道列标题的名称，这样它才能工作，而我无法了解如何使用字段值作为列标题。我试图做的是可能的，还是应该在以后用python手动构建该结构

很高兴能得到任何帮助

编辑：数据库引擎是Microsoft SQL

---

编辑：正在运行的解决方案：

您可以使用此查询

DECLARE @ColNames NVARCHAR(MAX) = ''
SELECT @ColNames = @ColNames + ', ' + QUOTENAME( DUN_ID ) FROM MyTable
GROUP BY DUN_ID

DECLARE @SqlText NVARCHAR(MAX) = '
SELECT * FROM (SELECT MEA_DateTime, PTR_ID, EXP_ID, DUN_ID, MEA_Value FROM MyTable ) SRC
PIVOT( MAX(MEA_Value) FOR DUN_ID IN ( '+ STUFF(@ColNames,1,1,'') +') ) AS PVT ORDER BY PTR_ID, EXP_ID'

EXEC(@SqlText)

结果:

MEA_DateTime            PTR_ID      EXP_ID      2          3        4       
----------------------- ----------- ----------- ---------- -------- --------
2009-08-10 00:00:00.000 24          14          15.10      14.30    16.70
2009-08-10 00:00:10.000 24          15          13.00      NULL     13.40
2009-08-10 00:00:20.000 24          16          17.80      17.70    16.20
2009-08-10 00:00:00.000 25          14          NULL       34.00    19.00
2009-08-10 00:00:10.000 25          15          22.10      23.10    NULL
2009-08-10 00:00:20.000 25          16          24.60      18.30    18.20

---

非常感谢。成功了！SQL FIDLE，以防有人要测试：