我在MYSQL中有一个表：

CREATE TABLE test.tem(a INT,b INT);    

以下数据：

INSERT INTO test.tem VALUES(1,2),(1,1),NULL),(2,3);

现在数据应该是：

+------+------+
| a    | b    |
+------+------+
|    1 |    2 |
|    1 |    1 |
|    1 | NULL |
|    2 |    3 |
+------+------+

我想按列a将列b更新到min(b)组.

所以SQL应该是：

UPDATE test.tem o
SET o.b = (SELECT
             MIN(b)
           FROM test.tem i
           WHERE i.a = o.a)

但是MYSQL无法在FROM子句中指定更新目标表

所以我认为下面的SQL可以很好地解决我的问题：

UPDATE test.tem t1
  JOIN test.tem t2
    ON t1.a = t2.a
SET t1.b = t2.b
WHERE t1.b IS NULL
     OR t1.b > t2.b;

但结果是：

+------+------+
| a    | b    |
+------+------+
|    1 |    1 |
|    1 |    1 |
|    1 |    2 |
|    2 |    3 |
+------+------+

其实我需要的结果是：

+------+------+
| a    | b    |
+------+------+
|    1 |    1 |
|    1 |    1 |
|    1 |    1 |
|    2 |    3 |
+------+------+

问题1：为什么MYSQL使用SQL计算出错误的结果？具有高效率的正确SQL应该是什么？
问题2：如果我只想用NULL值更新b(仅更新第三条记录),SQL应该是什么？

关于问题2,我试图使用下面不正确的SQL：

UPDATE test.tem t1
  JOIN test.tem t2
    ON t1.a = t2.a
    AND t1.b IS NULL
SET t1.b = t2.b
WHERE t1.b IS NULL
     OR t1.b > t2.b;

最佳答案
您没有唯一列来标识行.因此,yourJOIN可能会按您的意愿更新更多行.

你可能想要这样的东西：

UPDATE tem AS t1 JOIN (SELECT a,MIN(b) AS m FROM tem GROUP BY a) AS t2
USING (a)
SET t1.b = t2.m;

见http://sqlfiddle.com/#!2/c6a04/1

如果您只想更新列b中具有NULL的行,这只是WHERE子句的问题：

CREATE TABLE tem(a INT,b INT);    
INSERT INTO tem VALUES(1,3);

UPDATE tem AS t1 JOIN (SELECT a,MIN(b) AS m FROM tem GROUP BY a) AS t2
USING (a)
SET t1.b = t2.m
WHERE t1.b IS NULL;

见http://sqlfiddle.com/#!2/31ffb/1