UNION ALL
有两个表,A和B。
A表数据类似
id | a1 | a2 | a3 | a4 | a5 |
1 | 5 | 0.5 | a | g1 | ... |
2 | 10 | 1.0 | a | g2 | ... |
3 | 15 | 1.5 | a | g1 | ... |
B表数据类似
id | a1 | a2 | a3 | a4 | a5 |
1 | 10 | 2 | b | g2 | ... |
2 | 20 | 4 | b | g1 | ... |
3 | 30 | 6 | b | g1 | ... |
SELECT count(*), sum(x),sum(y),z,w FROM (
(SELECT a1 as x, a2 as y, a3 as z, a4 as w FROM A)
UNION ALL
(SELECT b1 as x, b3 as z, b2 as y, b4 as w FROM B)
)t GROUP BY w
结果发现,sum(x)结果是对的,但sum(y)是错的,为什么呢?
因为按照我的思路,UNION的select结果应该是:
x | y | z | w |
5 | 0.5 | a | g1 |
10 | 1.0 | a | g2 |
15 | 1.5 | a | g1 |
10 | 2 | b | g2 |
20 | 4 | b | g1 |
30 | 6 | b | g1 |
但实际上结果却是:
x | y | z | w |
5 | 0.5 | a | g1 |
10 | 1.0 | a | g2 |
15 | 1.5 | a | g1 |
10 | b | 2 | g2 |
20 | b | 4 | g1 |
30 | b | 6 | g1 |
注意:y下面的三行不是数字,而是字母。
这时来看上面被union的两个select的字段顺序,
第一个是:a1,a2,a3,a4,
第二个是:b1,b3,b2,b4,
虽然其对应的xyzw都没对应错,但是:
union查询时并不会按你为各字段起的别名进行一一对应,而是按你的select时的字段顺序进行对应的