使用distinctedjoin更新MySQL

我有两张桌子：

CREATE TABLE `tmp1` (
    `c1` VARCHAR(50) NULL DEFAULT NULL,
    `c2` VARCHAR(50) NULL DEFAULT NULL
);

CREATE TABLE `tmp2` (
    `c1` VARCHAR(50) NULL DEFAULT NULL,
    `c2` VARCHAR(50) NULL DEFAULT NULL
);

c2是主键。 对于tmp1，我有如下值：

|| 1 || val1 ||
|| 2 || val2 ||
|| 2 || val3 ||

对于tmp2：

|| 1 || val11 ||
|| 2 || val22 ||
|| 2 || val33 ||
|| 2 || val44 ||

我需要通过加入c1字段来更新tmp2，以接收如下结果：

|| 1 || val1 ||
|| 2 || val2 ||
|| 2 || val3 ||
|| 2 || val44 ||

|| 1 || val1 ||  ->(by c1)  || 1 || val11 ||
|| 2 || val2 ||  ->(by c1)  || 2 || val22 ||
|| 2 || val3 ||  ->(by c1)  || 2 || val33 ||

当我使用常规内部联接时，我将收到错误的结果：

|| 1 || val1 ||
|| 2 || val2 ||
|| 2 || *val2* ||
|| 2 || *val2* ||

我需要加入更新，如下所示：

|| 1 || val1 ||
|| 2 || val2 ||
|| 2 || val3 ||
|| 2 || val44 ||

|| 1 || val1 ||  ->(by c1)  || 1 || val11 ||
|| 2 || val2 ||  ->(by c1)  || 2 || val22 ||
|| 2 || val3 ||  ->(by c1)  || 2 || val33 ||

或者像这样，我想可能更真实

|| 1 || val1 ||  ->(by c1)  || 1 || val11 ||
|| 2 || val2 ||  ->(by c1)  || 2 || val22 ||

所以在第二种情况下，我认为我需要将tmp2按c1分组，将tmp1按c1分组，然后只更新该组中的一行

此选择返回我需要更新的内容：

select tmp1.c2, tmp2.c2 from tmp1
inner join tmp2 on tmp1.c1 = tmp2.c1 
group by tmp2.c1

|| val1 || val11 ||
|| val2 || val22 ||

---

我解决了此查询的问题：

UPDATE tmp2 as up
JOIN (select ij.c2 as x, src.c2 as y from tmp1 as src
inner join tmp2 as ij on src.c1 = ij.c1 
group by src.c1) as sel ON up.c2 = sel.x
SET up.c2 = y

---

当tmp2中的第一列相同时，它如何知道应该从tmp1复制哪一行？你需要一些区分行的方法。不清楚你想要什么。根据SQL标准，MySQL正在返回正确的结果，但是如果没有精确的查询，很难说更新一定是这样的。但这是错误的sql：更新tmp2为up set up.c2=从tmp1中选择tmp1.c2作为j在tmp1.c1=j.c1中，up.c1=tmp1.c1按j.c1分组这也是错误的查询=更新tmp2为up join从tmp1中选择ij.c2作为x，src.c2作为y在src.c1=ij.c1组中选择src.c1作为ij在up.c2=x set up.c2=y