尝试删除 SQL Server 中的重复行，其中差异是日期或批号

Posted 2023-03-29

【中文标题】尝试删除 SQL Server 中的重复行，其中差异是日期或批号

【英文标题】：Trying to delete duplicate rows in SQL Server where the difference is the date or batch number

【发布时间】：2021-12-25 07:22:26

【问题描述】：

我有这个问题：

SELECT 
    T1.ID_NUMBER,                                                                    
    T1.INCEPTION_DATE,
    T1.OCCURRENCE,
    T1.TRANSACTION_DATE,
    T1.FILE_LOAD_DATE,
    T1.BATCH_NUM
FROM 
    mastertable T1
INNER JOIN 
    (SELECT 
         ID_NUMBER, INCEPTION_DATE, OCCURRENCE, 
         COUNT(*) AS DUPL_COUNT
     FROM 
         mastertable
     WHERE 
         SOURCE_SYSTEM ='LEGACY'
     GROUP BY 
         ID_NUMBER, INCEPTION_DATE, OCCURRENCE
     HAVING 
         COUNT(*) > 1) t2 ON T2.ID_NUMBER = T1.ID_NUMBER 
                          AND T2.INCEPTION_DATE = T1.INCEPTION_DATE 
                          AND T2.OCCURRENCE= T1.OCCURRENCE
ORDER BY 
    1, 2, 3, 4, 5

返回以下结果

ID_NUMBER	INCEPTION_DATE	OCCURRENCE	TRANSACTION_DATE	FILE_LOAD_DATE	BATCH_NUM
112897732	2008-09-15	4	2008-07-03	2008-07-07 17:57:19	06341
112897732	2008-09-15	4	2008-07-13	2008-07-18 03:35:55	06753
828194721	2008-11-11	1	2008-09-06	2008-09-17 02:50:44	97334
828194721	2008-11-11	1	2008-09-23	2008-09-24 02:55:27	98331
456457422	2008-09-28	1	2008-12-03	2008-07-13 08:08:39	00734
456457422	2008-09-28	1	2008-12-03	2008-07-18 13:35:55	00991
999272910	2008-05-07	3	2008-05-03	2008-10-13 08:08:38	11432
999272910	2008-05-07	3	2008-05-28	2008-10-18 03:35:55	13342
875328642	2008-03-01	3	2008-04-28	2008-01-23 08:08:38	74542
875328642	2008-03-01	3	2008-04-30	2008-01-25 12:55:11	77536
011028734	2008-07-12	2	2008-12-03	2008-08-07 11:57:03	23422
011028734	2008-07-12	2	2008-12-03	2008-08-11 17:23:29	25748
018264981	2008-07-09	0	2008-12-03	2008-12-07 02:18:12	00432
018264981	2008-07-09	0	2008-12-03	2008-12-11 17:44:19	00773

每个ID_NUMBER或更早的FILE_LOAD_DATE或更小的BATCH_NUM是我要保留的记录。

有没有办法编写一个删除其他记录的查询，也许使用带有ROW_NUMBER() 的 CTE？

我希望有一些干燥的东西，以防这个问题再次发生。谢谢！

（如果不是太麻烦，请解释解决方案的工作原理）

【参考方案1】：

您可以在此处使用可删除的 CTE：

WITH cte AS (
    SELECT *, ROW_NUMBER() OVER (PARTITION BY ID_NUMBER, INCEPTION_DATE, OCCURRENCE
                                 ORDER BY FILE_LOAD_DATE, BATCH_NUM) rn
    FROM mastertable
    WHERE SOURCE_SYSTEM = 'LEGACY'
)

DELETE
FROM cte
WHERE rn > 1;

逻辑是为ID_NUMBER、INCEPTION_DATE 和OCCURRENCE 具有相同值的每组记录分配一个行号。第一行编号值 1 将分配给具有 最早 FILE_LOAD_DATE 的记录。如果有两个或多个记录与最早的FILE_LOAD_DATE 并列，则决胜局将由最早的BATCH_NUM 确定。

delete 语句删除所有记录除了最早的记录。