以下示例均是在hive中进行操作
创建表 t1：
create table a(id int,name string)
row format delimited
fields terminated by ',';
构造数据：
1,zs
2,ls
3,ww
加载数据到表t1：
load data local inpath '/root/a.dat' into table a;
创建表 t2：
create table b(id int,nickname string)
row format delimited
fields terminated by ',';
构造数据：
1,zzz
3,www
4,sss
将数据加载到表t2：
load data local inpath '/root/b.dat' into table b;
--两表的数据：
0: jdbc:hive2://cts03:10000> select * from a;
+-------+---------+--+
| a.id | a.name |
+-------+---------+--+
| 1 | zs |
| 2 | ls |
| 3 | ww |
+-------+---------+--+
0: jdbc:hive2://cts03:10000> select * from b;
+-------+-------------+--+
| b.id | b.nickname |
+-------+-------------+--+
| 1 | zzz |
| 3 | www |
| 4 | sss |
+-------+-------------+--+
-- 笛卡尔积：
0: jdbc:hive2://cts03:10000> select a.*,b.* from a join b;
+-------+---------+-------+-------------+--+
| a.id | a.name | b.id | b.nickname |
+-------+---------+-------+-------------+--+
| 1 | zs | 1 | zzz |
| 2 | ls | 1 | zzz |
| 3 | ww | 1 | zzz |
| 1 | zs | 3 | www |
| 2 | ls | 3 | www |
| 3 | ww | 3 | www |
| 1 | zs | 4 | sss |
| 2 | ls | 4 | sss |
| 3 | ww | 4 | sss |
+-------+---------+-------+-------------+--+
--内连接：
0: jdbc:hive2://cts03:10000> select a.*,b.* from a join b on a.id=b.id;
+-------+---------+-------+-------------+--+
| a.id | a.name | b.id | b.nickname |
+-------+---------+-------+-------------+--+
| 1 | zs | 1 | zzz |
| 3 | ww | 3 | www |
+-------+---------+-------+-------------+--+
--左外连接：左表数据全返回给我作为查询的输入数据集
0: jdbc:hive2://cts03:10000> select a.*,b.* from a left join b on a.id=b.id;
+-------+---------+-------+-------------+--+
| a.id | a.name | b.id | b.nickname |
+-------+---------+-------+-------------+--+
| 1 | zs | 1 | zzz |
| 2 | ls | NULL | NULL |
| 3 | ww | 3 | www |
+-------+---------+-------+-------------+--+
--右外连接：右表的数据劝返回作为查询的输入数据集
0: jdbc:hive2://cts03:10000> select a.*,b.* from a right join b on a.id=b.id;
+-------+---------+-------+-------------+--+
| a.id | a.name | b.id | b.nickname |
+-------+---------+-------+-------------+--+
| 1 | zs | 1 | zzz |
| 3 | ww | 3 | www |
| NULL | NULL | 4 | sss |
+-------+---------+-------+-------------+--+
--全外连接：两表的数据都需要返回作为查询的输入数据集
0: jdbc:hive2://cts03:10000> select a.*,b.* from a full join b on a.id=b.id;
+-------+---------+-------+-------------+--+
| a.id | a.name | b.id | b.nickname |
+-------+---------+-------+-------------+--+
| 1 | zs | 1 | zzz |
| 2 | ls | NULL | NULL |
| 3 | ww | 3 | www |
| NULL | NULL | 4 | sss |
+-------+---------+-------+-------------+--+
--左半连接（hive特有）：按照内连接的规律连，但是只返回左半部分作为查询的输入数据集
0: jdbc:hive2://cts03:10000> select a.* from a left semi join b on a.id=b.id;
+-------+---------+
| a.id | a.name |
+-------+---------+
| 1 | zs |
| 3 | ww |
+-------+---------+
注： 左半连接是在mapreduce这种运算模型下对in子句的更高效实现
select id,name from a where id in(select distinct id from b)