Oracle SQL-将列转换为行分组

我在Oracle 11g实例上运行了以下SQL查询，效果非常好：

SELECT
    TRIM(PART_NM) || TRIM(PART_NO),
    TRIM(MKBY_IND),
    MFG_TYP

FROM
    TblPARTS

WHERE
    ORDER_START_DATE >= (SYSDATE - 365)
    AND
    TO_CHAR(TRIM(PART_NO)) != '99'

这将产生如下结果：

| Part | Make / Buy |    Area   |
|:----:|:----------:|:---------:|
| AU54 |      M     |    Demo   |
| ZA45 |      M     | Construct |
| AD28 |      B     | Construct |
| SX73 |      B     | Construct |
| MO47 |      M     | Construct |
| IN66 |      B     | Construct |
| RP42 |      M     |    Demo   |
| HF77 |      B     |    Demo   |
| IE43 |      B     |  Personal |
| OX11 |      B     |  Personal |
| ZE45 |      M     |    Demo   |
| JH21 |      M     | Construct |
| AM91 |      M     | Construct |
| DS50 |      M     |    Demo   |
| LE29 |      M     |    Demo   |
| IP91 |      M     |    Demo   |
| ID42 |      M     |    Demo   |
| RI96 |      M     |    Demo   |

我在那里没有问题。但是，有人要求我转置Area列，而不是按它对行进行分组。类似于：

|    Part   | Make / Buy |
|:---------:|:----------:|
| CONSTRUCT |  CONSTRUCT |
|    AU54   |      M     |
|    ZA45   |      M     |
|    AD28   |      B     |
|    SX73   |      M     |
|    MO47   |      M     |
|    IN66   |      B     |
|    JH21   |      M     |
|    AM91   |      B     |
|    DEMO   |    DEMO    |
|    RP42   |      M     |
|    HF77   |      M     |
|    ZE45   |      M     |
|    DS50   |      B     |
|    LE29   |      M     |
|    IP91   |      M     |
|    ID42   |      M     |
|    RI96   |      M     |
|  PERSONAL |  PERSONAL  |
|    IE43   |      M     |
|    OX11   |      M     |

简言之，我希望在“区域”列（

MFG_TYP

）字段中找到唯一的值，并能够按它们按行分组。应该注意的是，列出的三个条目在我的数据集中是有效的，但可能会发生更改，因此我不是在寻找静态解决方案，而是在

区域

列中的所有唯一值上动态应用转换的方法

我曾尝试查看这些资源，但目前不了解如何构建解决方案。任何建议都将不胜感激

---

看起来这就是你要找的：

SELECT
    'CONSTRUCT' AS "Part",
    'CONSTRUCT' AS "Make / Buy"
FROM DUAL

UNION

SELECT
    TRIM(PART_NM) || TRIM(PART_NO) AS "Part",
    TRIM(MKBY_IND) AS "Make / Buy"
FROM
    TblPARTS
WHERE
    ORDER_START_DATE >= (SYSDATE - 365)
    AND    TO_CHAR(TRIM(PART_NO)) != '99'
    AND    MFG_TYP = 'Construct'

UNION

SELECT
    'DEMO' AS "Part",
    'DEMO' AS "Make / Buy"
FROM DUAL

UNION

SELECT
    TRIM(PART_NM) || TRIM(PART_NO) AS "Part",
    TRIM(MKBY_IND) AS "Make / Buy"
FROM
    TblPARTS
WHERE
    ORDER_START_DATE >= (SYSDATE - 365)
    AND    TO_CHAR(TRIM(PART_NO)) != '99'
    AND    MFG_TYP = 'Demo'

UNION

SELECT
    'PERSONAL ' AS "Part",
    'PERSONAL ' AS "Make / Buy"
FROM DUAL

UNION

SELECT
    TRIM(PART_NM) || TRIM(PART_NO)) AS "Part",
    TRIM(MKBY_IND) AS "Make / Buy"
FROM
    TblPARTS
WHERE
    ORDER_START_DATE >= (SYSDATE - 365)
    AND    TO_CHAR(TRIM(PART_NO)) != '99'
    AND    MFG_TYP = 'Personal'

快速解决方案包括：

---

在线测试。

请提供您的样本数据，这将对您有所帮助。样本数据包含在帖子中。哪些内容需要编辑@AviI只看到您从查询中获得的结果和预期结果。是。这不符合样本数据的条件吗@阿维。我可以从表中选择*，但我不知道这会有多大好处。我很乐意为您提供您需要的帮助，但在过去，我也被告知使用我目前拥有的，然后使用我的预期结果。如果不仅仅是这三个

MFG_TYP

字段，这怎么可能起作用？也就是说，这个数字是动态的。我怎么可能对这三个变量进行静态调整，而不是对列中的所有唯一值进行静态调整？你能解释一下

<7

，这是为什么吗？@JerryM.，

7

是位向量

111

的十进制表示形式，它对应于一个超集合行，其中

区域

、

部分

和

购买

列的值为

空

。看看

with
  TblPARTS(Part, Make_Buy, Area) as (
    select 'AU54', 'M', 'Demo' from dual union all
    select 'ZA45', 'M', 'Construct' from dual union all
    select 'AD28', 'B', 'Construct' from dual union all
    select 'SX73', 'B', 'Construct' from dual union all
    select 'MO47', 'M', 'Construct' from dual union all
    select 'IN66', 'B', 'Construct' from dual union all
    select 'RP42', 'M', 'Demo' from dual union all
    select 'HF77', 'B', 'Demo' from dual union all
    select 'IE43', 'B', 'Personal' from dual union all
    select 'OX11', 'B', 'Personal' from dual union all
    select 'ZE45', 'M', 'Demo' from dual union all
    select 'JH21', 'M', 'Construct' from dual union all
    select 'AM91', 'M', 'Construct' from dual union all
    select 'DS50', 'M', 'Demo' from dual union all
    select 'LE29', 'M', 'Demo' from dual union all
    select 'IP91', 'M', 'Demo' from dual union all
    select 'ID42', 'M', 'Demo' from dual union all
    select 'RI96', 'M', 'Demo' from dual union all
    select 'AO42', 'M', 'Public' from dual union all
    select 'OA42', 'B', 'Public' from dual
  ),
  a as (
    select
      Part, Make_Buy, Area,
      GROUPING_ID(Area, Part, Make_Buy) gid
    from TblPARTS
    group by rollup(Area, (Part, Make_Buy))
    having GROUPING_ID(Area, Part, Make_Buy) < 7
  )
select
  coalesce(Part, Area) Part,
  coalesce(Make_Buy, Area) Make_Buy
from a
order by Area, gid desc

+------------+-----------+
|    PART    |  MAKE_BUY |
+------------+-----------+
| Construct  | Construct |
| AM91       | M         |
| IN66       | B         |
| JH21       | M         |
| SX73       | B         |
| ZA45       | M         |
| AD28       | B         |
| MO47       | M         |
| Demo       | Demo      |
| ZE45       | M         |
| RP42       | M         |
| RI96       | M         |
| LE29       | M         |
| IP91       | M         |
| ID42       | M         |
| HF77       | B         |
| AU54       | M         |
| DS50       | M         |
| Personal   | Personal  |
| OX11       | B         |
| IE43       | B         |
| Public     | Public    |
| OA42       | B         |
| AO42       | M         |
+------------+-----------+