删除除一组最大值之外的所有记录

Question

我有这张桌子：

CREATE TABLE tblTest
(
    Id INT PRIMARY KEY NOT NULL,
    Weight INT NULL,
    Batch INT NULL,
    t_stamp DATETIME NULL
);

让我们假设它充满了价值观：

INSERT INTO BatchDataWillieNelson (WT401_Raw, Batch_Done_Signal,
                                   Batch_Number, t_stamp)
VALUES (31, 3994, 2, '2023-12-05 06:19:58.990'),
       (32, 4052, 2, '2023-12-05 06:37:24.883'),
       (33, 4000, 5, '2023-12-05 08:44:09.780'),
       (34, 4058, 5, '2023-12-05 08:58:59.683'),
       (35, 4032, 8, '2023-12-05 11:18:11.727'),
       (36, 3983, 11, '2023-12-05 13:20:35.167'),
       (37, 4013, 11, '2023-12-05 13:33:54.877'),
       (38, 3993, 14, '2023-12-05 15:29:50.060'),
       (39, 3470, 14, '2023-12-05 15:48:51.053'),
       (40, 3996, 14, '2023-12-05 17:20:09.893'),
       (41, 3348, 14, '2023-12-05 17:39:22.477'),
       (42, 3993, 20, '2023-12-05 19:48:42.117'),
       (43, 4054, 20, '2023-12-05 20:09:12.217'),
       (44, 3991, 23, '2023-12-05 22:55:45.020'),
       (45, 4065, 23, '2023-12-05 23:16:12.077'),
       (46, 3993, 26, '2023-12-06 02:56:29.713'),
       (47, 4033, 26, '2023-12-06 03:12:24.483'),
       (48, 3982, 26, '2023-12-06 06:06:45.720'),
       (49, 4075, 26, '2023-12-06 06:35:05.197'),
       (50, 3994, 4, '2023-12-06 08:02:58.887'),
       (51, 3029, 4, '2023-12-06 08:28:21.260'),
       (52, 3982, 7, '2023-12-06 09:49:23.497'),
       (53, 4038, 7, '2023-12-06 10:12:41.707')

我正在使用 SQL Server 2008。我有一个表，其中有错误写入的额外记录。我需要删除这些。

我需要根据

Weight

的分组来评估

Batch_Number

列的最大值。同一个

Batch_Number

可以多次使用。因此，在没有对表进行排序的情况下，我尝试查找

Batch_Number

中的更改，并将这些行视为一个组来进行评估，而不必遍历每条记录。

预期结果：

身份证	重量	批量	t_邮票
32	4052	2	'2023-12-05 06:37:24.883'
34	4058	5	'2023-12-05 08:58:59.683'
35	4032	8	'2023-12-05 11:18:11.727'
37	4013	11	'2023-12-05 13:33:54.877'
40	3996	14	'2023-12-05 17:20:09.893'
43	4054	20	'2023-12-05 20:09:12.217'
49	4075	26	'2023-12-06 06:35:05.197'
50	3994	4	'2023-12-06 08:02:58.887'
53	4038	7	'2023-12-06 10:12:41.707'

Answer 1

WITH MaxWeightPerBatch AS (
    SELECT Batch_Number, MAX(Weight) AS MaxWeight
    FROM tblTest
    GROUP BY Batch_Number
)
DELETE FROM tblTest
WHERE (Batch_Number, Weight) NOT IN (
    SELECT Batch_Number, MaxWeight
    FROM MaxWeightPerBatch
);

找到每个 Batch_Number 组的最大权重，然后从 tblTest 中删除在各自的 Batch_Number 组内没有最大权重的行。

Answer 2

删除存在具有相同批次号和更高权重/信号的另一行的所有行。

delete from batchdatawillienelson
where exists
(
  select null
  from batchdatawillienelson other
  where other.batch_number = batchdatawillienelson.batch_number
  and other.batch_done_signal > batchdatawillienelson.batch_done_signal
);

Answer 3

如果您的 RDBMS 支持直接对表表达式进行操作，例如 SQL Server，则以下是另一种简单的方法：

delete from d
from (
    select *, Row_Number() over(partition by batch order by weight desc) rn
    from t
)d
where rn > 1;

参见 Fiddle 演示