排名:如何将 ROW_NUMBER 或 RANK 重置为 1

Ranking: How to reset the ROW_NUMBER or RANK to 1

使用 SQL 服务器 2014:

考虑以下 table:

DECLARE @Table TABLE (
      Id int NOT NULL identity(1,1),
      Col_Value varchar(2)
)

INSERT INTO @Table (Col_Value)
VALUES ('A'),('A'),('B'),('B'),('B'),('A'),('A'),('B'),('B'),('B'),('A'),('B'),('B'),('A'),('A'),('B'),('C'),('C'),('A'),('A'),('B'),('B'),('C')

如何创建一个查询,在结果中生成 R 列,如下所示

+----+------+---+
| ID | Data | R |
+----+------+---+
| 1  | A    | 1 |
+----+------+---+
| 2  | A    | 2 |
+----+------+---+
| 3  | B    | 1 |
+----+------+---+
| 4  | B    | 2 |
+----+------+---+
| 5  | B    | 3 |
+----+------+---+
| 6  | A    | 1 |
+----+------+---+
| 7  | A    | 2 |
+----+------+---+
| 8  | B    | 1 |
+----+------+---+
| 9  | B    | 2 |
+----+------+---+
| 10 | B    | 3 |
+----+------+---+
| 11 | A    | 1 |
+----+------+---+
| 12 | B    | 1 |
+----+------+---+
| 13 | B    | 2 |
+----+------+---+
| 14 | A    | 1 |
+----+------+---+
| 15 | A    | 2 |
+----+------+---+
| 16 | B    | 1 |
+----+------+---+
| 17 | C    | 1 |
+----+------+---+
| 18 | C    | 2 |
+----+------+---+
| 19 | A    | 1 |
+----+------+---+
| 20 | A    | 2 |
+----+------+---+
| 21 | B    | 1 |
+----+------+---+
| 22 | B    | 2 |
+----+------+---+
| 23 | C    | 1 |
+----+------+---+

在上面的结果table中,一旦数据列在一行中发生变化,R值就会重置为1

更新 1

Ben Thul 的回答非常有效。

我建议在下面 post 更新参考这个答案。

T-sql Reset Row number on Field Change

这在文献中称为 "gaps and islands" 问题。首先,我提出的解决方案:

with cte as (
    select *, [Id] - row_number() over (partition by [Col_Value] order by [Id]) as [GroupID]
    from @table
)
select [Id], [Col_Value], row_number() over (partition by [GroupID], [Col_Value] order by [Id])
from cte
order by [Id];

为了说明,请注意,如果我使用 row_number() 枚举所有 "A" 值,那些连续的 row_number() 值将以与 ID 相同的速率上升价值。也就是说,他们的差异对于那个连续组(也称为 "island")中的那些人来说是相同的。一旦我们计算出该组标识符,只需枚举每个组的每个成员即可。