一條SQL刪除重復記錄，重復的只保留一條

 

情景：

　　我們的數據庫中可能會存在很多因各種原因而重復的記錄，我們需要對這些重復的記錄進行刪除，每組組重復的記錄只保留一條就行

例如我們有這么個表：兩個框框都是有重復記錄的，紅框和綠框都只需要留下一條，其他的都干掉。

 

 

前提：

　　1：每天記錄都要有一個唯一id

　　2：每組重復的記錄要有字段能進行分組，例如上面我們按name、class、age、score相同的就是一組。

數據寶貴，請先備份！！！

數據寶貴，請先備份！！！

數據寶貴，請先備份！！！

放上整條SQL：

DELETE FROM "t_hw_test_del" t3
WHERE 1 = 1 AND EXISTS (
    SELECT * FROM (
        SELECT
             "id","name","class","age","score"
        FROM
            "t_hw_test_del"
        WHERE 1 = 1
    )t1 LEFT JOIN (
        SELECT
            "MAX"("id") as "id","name","class","age","score"
        FROM
            "t_hw_test_del"
        WHERE 1 = 1
        GROUP BY "name","class","age","score"
    ) t2 ON t1."id" = t2."id" 
    AND t1."name" = t2."name" 
    AND t1."class" = t2."class" 
    AND t1."age" = t2."age" 
    AND t1."score" = t2."score"
    WHERE t2."id" is NULL AND t1."id" = t3."id"
) 

 

 

按步分析：

1：找到我們需要的記錄，因為我們可能只是要處理某天的記錄，所以要先篩選一下，我這里用1=1來代替

SELECT
    *
FROM
    "t_hw_test_del"
WHERE 1 = 1

 

 

 

2：找到我們每一組要保留的記錄 ，我這里選每組id最大的保留，其他的不要

SELECT
    "MAX"("id") as "id","name","class","age","score"
FROM
    "t_hw_test_del"
WHERE 1 = 1
GROUP BY "name","class","age","score"

 

 

3：用我們找到的需要處理的記錄和要保留的記錄關聯起來, 用分組字段和唯一id關聯，左聯

SELECT * FROM (
    SELECT
         "id","name","class","age","score"
    FROM
        "t_hw_test_del"
    WHERE 1 = 1
)t1 LEFT JOIN (
    SELECT
        "MAX"("id") as "id","name","class","age","score"
    FROM
        "t_hw_test_del"
    WHERE 1 = 1
    GROUP BY "name","class","age","score"
) t2 ON t1."id" = t2."id" 
AND t1."name" = t2."name" 
AND t1."class" = t2."class" 
AND t1."age" = t2."age" 
AND t1."score" = t2."score"

 

 

4：右表為null的記錄，對應的左表記錄就是我們要刪掉的，加個條件右表為null

SELECT * FROM (
    SELECT
         "id","name","class","age","score"
    FROM
        "t_hw_test_del"
    WHERE 1 = 1
)t1 LEFT JOIN (
    SELECT
        "MAX"("id") as "id","name","class","age","score"
    FROM
        "t_hw_test_del"
    WHERE 1 = 1
    GROUP BY "name","class","age","score"
) t2 ON t1."id" = t2."id" 
AND t1."name" = t2."name" 
AND t1."class" = t2."class" 
AND t1."age" = t2."age" 
AND t1."score" = t2."score"
WHERE t2."id" is NULL

 

 

5： 用我們原來的記錄id在我們不需要的記錄里面找，如果這條記錄在我們不需要的記錄集里，那這條記錄就可以刪除

DELETE FROM "t_hw_test_del" t3
WHERE 1 = 1 AND EXISTS (
    SELECT * FROM (
        SELECT
             "id","name","class","age","score"
        FROM
            "t_hw_test_del"
        WHERE 1 = 1
    )t1 LEFT JOIN (
        SELECT
            "MAX"("id") as "id","name","class","age","score"
        FROM
            "t_hw_test_del"
        WHERE 1 = 1
        GROUP BY "name","class","age","score"
    ) t2 ON t1."id" = t2."id" 
    AND t1."name" = t2."name" 
    AND t1."class" = t2."class" 
    AND t1."age" = t2."age" 
    AND t1."score" = t2."score"
    WHERE t2."id" is NULL AND t1."id" = t3."id"
) 

 

 6： 檢查一下結果符不符合我們的要求

SELECT
     "id","name","class","age","score"
FROM
    "t_hw_test_del"
WHERE 1 = 1