序列化表格详细信息

Question

我正在使用SQL Server 2012，我想使用更新语句重新排列下面的订单号，但为每个订单行保留相同的订单号

order_date  Order_Number  Order_Line
-------------------------------------
2015-01-06      1               1
2015-01-06      1               2
2015-01-06      6               1
2015-01-06      6               2
2015-01-06      9               1
2015-01-06      9               2
2015-01-06      9               3
2015-01-06      11              1
2015-01-06      11              2
2015-01-06      11              3
2015-01-06      11              4

我试过这个：

DECLARE @OrdN INT = 0 

UPDATE hist_line 
SET @OrdN = order_number = @OrdN + 1 
WHERE order_date = '06/01/2015'

但它为每一行返回一个新的序列，它应该如下所示：

order_date  Order_Number  Order_Line
-------------------------------------
2015-01-06      1               1
2015-01-06      1               2
2015-01-06      2               1
2015-01-06      2               2
2015-01-06      3               1
2015-01-06      3               2
2015-01-06      3               3
2015-01-06      4               1
2015-01-06      4               2
2015-01-06      4               3
2015-01-06      4               4

Answer 1

您可以使用dense_rank()：

select order_date,
       dense_rank() over (partition by order_date order by Order_Number) as Order_Number,
       Order_Line
from t;

这是一个ANSI标准函数，可在大多数数据库中使用。

编辑：

要实际执行更新，请使用可更新的CTE：

with toupdate as (
      select t.*,
             dense_rank() over (partition by order_date order by Order_Number) as new_Order_Number
      from t
     )
update toupdate
    set Order_Number = new_Order_Number;