Union和Union All的区别

假设我们有一个表Student，包括以下字段与数据：

[c-sharp] view plain copy print ?

1. drop table student;
3. create table student
4. (
5. id int primary key,
6. name nvarchar2(50) not null,
7. score number not null
8. );
10. insert into student values(1,'Aaron',78);
11. insert into student values(2,'Bill',76);
12. insert into student values(3,'Cindy',89);
13. insert into student values(4,'Damon',90);
14. insert into student values(5,'Ella',73);
15. insert into student values(6,'Frado',61);
16. insert into student values(7,'Gill',99);
17. insert into student values(8,'Hellen',56);
18. insert into student values(9,'Ivan',93);
19. insert into student values(10,'Jay',90);
21. commit;

drop table student; create table student ( id int primary key, name nvarchar2(50) not null, score number not null ); insert into student values(1,'Aaron',78); insert into student values(2,'Bill',76); insert into student values(3,'Cindy',89); insert into student values(4,'Damon',90); insert into student values(5,'Ella',73); insert into student values(6,'Frado',61); insert into student values(7,'Gill',99); insert into student values(8,'Hellen',56); insert into student values(9,'Ivan',93); insert into student values(10,'Jay',90); commit;

首先，我们来看一下UNION的例子：

[c-sharp] view plain copy print ?

1. SQL> select *
2. 2 from student
3. 3 where id<4
4. 4 union
5. 5 select *
6. 6 from student
7. 7 where id>2 and id<6
8. 8 ;
10. ID NAME SCORE
11. ---------- ------------------------------ ----------
12. 1 Aaron 78
13. 2 Bill 76
14. 3 Cindy 89
15. 4 Damon 90
16. 5 Ella 73
18. SQL>

SQL> select * 2 from student 3 where id<4 4 union 5 select * 6 from student 7 where id>2 and id<6 8 ; ID NAME SCORE ---------- ------------------------------ ---------- 1 Aaron 78 2 Bill 76 3 Cindy 89 4 Damon 90 5 Ella 73 SQL>

如果换成Union All连接两个结果集，则结果如下：

[c-sharp] view plain copy print ?

1. SQL> select *
2. 2 from student
3. 3 where id<4
4. 4 union all
5. 5 select *
6. 6 from student
7. 7 where id>2 and id<6
8. 8 ;
10. ID NAME SCORE
11. ---------- ------------------------------ ----------
12. 1 Aaron 78
13. 2 Bill 76
14. 3 Cindy 89
15. 3 Cindy 89
16. 4 Damon 90
17. 5 Ella 73
19. 6 rows selected.

SQL> select * 2 from student 3 where id<4 4 union all 5 select * 6 from student 7 where id>2 and id<6 8 ; ID NAME SCORE ---------- ------------------------------ ---------- 1 Aaron 78 2 Bill 76 3 Cindy 89 3 Cindy 89 4 Damon 90 5 Ella 73 6 rows selected.

可以看到，Union和Union All的区别之一在于对重复结果的处理。

接下来，我们交换一个两个SELECT语句的顺序，看看结果是怎样的。

[c-sharp] view plain copy print ?

1. SQL> select *
2. 2 from student
3. 3 where id>2 and id<6
4. 4 union
5. 5 select *
6. 6 from student
7. 7 where id<4
8. 8 ;
10. ID NAME SCORE
11. ---------- ------------------------------ ----------
12. 1 Aaron 78
13. 2 Bill 76
14. 3 Cindy 89
15. 4 Damon 90
16. 5 Ella 73
18. SQL> select *
19. 2 from student
20. 3 where id>2 and id<6
21. 4 union all
22. 5 select *
23. 6 from student
24. 7 where id<4
25. 8 ;
27. ID NAME SCORE
28. ---------- ------------------------------ ----------
29. 3 Cindy 89
30. 4 Damon 90
31. 5 Ella 73
32. 1 Aaron 78
33. 2 Bill 76
34. 3 Cindy 89
36. 6 rows selected.

SQL> select * 2 from student 3 where id>2 and id<6 4 union 5 select * 6 from student 7 where id<4 8 ; ID NAME SCORE ---------- ------------------------------ ---------- 1 Aaron 78 2 Bill 76 3 Cindy 89 4 Damon 90 5 Ella 73 SQL> select * 2 from student 3 where id>2 and id<6 4 union all 5 select * 6 from student 7 where id<4 8 ; ID NAME SCORE ---------- ------------------------------ ---------- 3 Cindy 89 4 Damon 90 5 Ella 73 1 Aaron 78 2 Bill 76 3 Cindy 89 6 rows selected.

可以看到，对于UNION来说，交换两个SELECT语句的顺序后结果仍然是一样的，这是因为UNION会自动排序。而UNION ALL在交换了SELECT语句的顺序后结果则不相同，因为UNION ALL不会对结果自动进行排序。

那么这个自动排序的规则是什么呢？我们交换一下SELECT后面选择字段的顺序（前面使用SELECT *相当于SELECT ID,NAME,SCORE），看看结果如何：

[c-sharp] view plain copy print ?

1. SQL> select score,id,name
2. 2 from student
3. 3 where id<4
4. 4 union
5. 5 select score,id,name
6. 6 from student
7. 7 where id>2 and id<6
8. 8 ;
10. SCORE ID NAME
11. ---------- ---------- ------------------------------
12. 73 5 Ella
13. 76 2 Bill
14. 78 1 Aaron
15. 89 3 Cindy
16. 90 4 Damon

SQL> select score,id,name 2 from student 3 where id<4 4 union 5 select score,id,name 6 from student 7 where id>2 and id<6 8 ; SCORE ID NAME ---------- ---------- ------------------------------ 73 5 Ella 76 2 Bill 78 1 Aaron 89 3 Cindy 90 4 Damon

可是看到，此时是按照字段SCORE来对结果进行排序的（前面SELECT *的时候是按照ID进行排序的）。

那么有人会问，如果我想自行控制排序，能不能使用ORDER BY呢？当然可以。不过在写法上有需要注意的地方：

[c-sharp] view plain copy print ?

1. select score,id,name
2. from student
3. where id > 2 and id < 7
5. union
7. select score,id,name
8. from student
9. where id < 4
11. union
13. select score,id,name
14. from student
15. where id > 8
16. order by id desc

select score,id,name from student where id > 2 and id < 7 union select score,id,name from student where id < 4 union select score,id,name from student where id > 8 order by id desc

order by子句必须写在最后一个结果集里，并且其排序规则将改变操作后的排序结果。对于Union、Union All、Intersect、Minus都有效。

其他的集合操作符，如Intersect和Minus的操作和Union基本一致，这里一起总结一下：

Union，对两个结果集进行并集操作，不包括重复行，同时进行默认规则的排序；

Union All，对两个结果集进行并集操作，包括重复行，不进行排序；

Intersect，对两个结果集进行交集操作，不包括重复行，同时进行默认规则的排序；

Minus，对两个结果集进行差操作，不包括重复行，同时进行默认规则的排序。

可以在最后一个结果集中指定Order by子句改变排序方式。