相同的 Row_Number 表示相同的值

Question

我需要进行行编号，其中 ROW_NUMBER 对于相同值列是相同的：MFGPN（相同的 MFGPN 将始终按顺序排列）。我还需要按No维持原来的顺序

这是我的桌子

No     MFGPN
1      Z363700Z01
2      Z363700Z01
3      0119-960-1
4      1A3F1-0503-01

我尝试使用 RANK() 来实现所需的目标，但遇到了麻烦。

SELECT RANK() OVER(ORDER BY MFGPN) As [Item], MFGPN FROM Table1 ORDER BY [No] ASC

结果

Item   MFGPN           Desired Result
3      Z363700Z01            1
3      Z363700Z01            1
1      0119-960-1            2
2      1A3F1-0503-01         3

感谢各位专家的建议。谢谢！

Answer 1

使用

DENSE_RANK

窗函数代替

RANK

。当存在重复数据时，

Rank

将跳过序列，

Dense_Rank

则不会。

SELECT MFGPN,
        Dense_rank()OVER(ORDER BY m_no) as [Desired Result]
FROM   (SELECT no,
                MFGPN,
                Min(no)OVER(partition BY MFGPN) AS m_no
        FROM   (VALUES (1,'Z363700Z01' ),
                        (2,'Z363700Z01' ),
                        (3,'0119-960-1' ),
                        (4,'1A3F1-0503-01')) tc (no, MFGPN))a 

如果

no

不唯一，则将

DENSE_RANK

更改为

Dense_rank()OVER(ORDER BY m_no,MFGPN)

结果：

+---------------+----------------+
|     MFGPN     | Desired Result |
+---------------+----------------+
| Z363700Z01    |              1 |
| Z363700Z01    |              1 |
| 0119-960-1    |              2 |
| 1A3F1-0503-01 |              3 |
+---------------+----------------+

Answer 2

您应该按

mfgpn

对结果进行分区，以便具有相同

mfgpn

的行按

no

获得相同的排名和顺序。此外，使用

dense_rank

将确保您不会“跳过”任何排名：

SELECT   DENSE_RANK() OVER(PARTITION BY [mfgpn] ORDER BY [no]) As [Item], 
         [mfgpm] 
FROM     Table1 
ORDER BY [No] ASC

Answer 3

select      sum(case when MFGPN = prev_MFGPN then 0 else 1 end) over (order by No)  as item
           ,MFGPN

from       (SELECT      lag(MFGPN) over (order by [No]) as prev_MFGPN   
                       ,[No]
                       ,MFGPN 
            FROM        Table1 
            ) t

ORDER BY    [No] ASC

---

+------+---------------+
| item | MFGPN         |
+------+---------------+
| 1    | Z363700Z01    |
+------+---------------+
| 1    | Z363700Z01    |
+------+---------------+
| 2    | 0119-960-1    |
+------+---------------+
| 3    | 1A3F1-0503-01 |
+------+---------------+

Answer 4

您可以尝试下面的代码片段。 DENSE_RANK() 上面解释的是这种情况下的最佳方法。

    SELECT a.*,
      DENSE_RANK() OVER(ORDER BY MFGPN DESC) RN
    FROM
      (SELECT 1 AS no, 'Z363700Z01' AS mfgpn FROM dual
      UNION ALL
      SELECT 2 AS no, 'Z363700Z01' AS mfgpn FROM dual
      UNION ALL
      SELECT 3 AS no, '0119-960-1' AS mfgpn FROM dual
      UNION ALL
      SELECT 4 AS no, '1A3F1-0503-01' AS MFGPN FROM dual
      )a;

-------------------------------OUTPUT-------------------------------------------
NO  MFGPN           RN
1   Z363700Z01      1
2   Z363700Z01      1
4   1A3F1-0503-01   2
3   0119-960-1      3

--------------------------------------------------------------------------------

Answer 5

尝试：我认为使用

join

并比较值以获得所需的输出更简单和灵活，如下所示：

SELECT p.MFGPN,
       Dense_rank()OVER(ORDER BY CASE WHEN p.MFGPN = tp.MFGPN THEN tp.num ELSE p.num END) AS [Desired Result]
FROM tmp_option p
LEFT JOIN tmp_option tp ON tp.num+1 = p.num
ORDER BY p.num ASC