Hive的left join、left outer join和left semi join三者的区别

2019-02-14 10:24:31 浏览数 (1)

Hive的Join的文档说明地址: https://cwiki.apache.org/confluence/display/Hive/LanguageManual+Joins

以下为两个测试数据表建表语句:

MySQL use test; DROP TABLE IF EXISTS table1; create table table1( student_no bigint comment '学号', student_name string comment '姓名' ) COMMENT 'test 学生信息' ROW FORMAT DELIMITED FIELDS TERMINATED BY 't' LINES TERMINATED BY 'n' STORED AS TEXTFILE; DROP TABLE IF EXISTS table2; create table table2( student_no bigint comment '学号', class_no bigint comment '课程号' ) COMMENT 'test 学生选课信息' ROW FORMAT DELIMITED FIELDS TERMINATED BY 't' LINES TERMINATED BY 'n' STORED AS TEXTFILE; load data local inpath 'data_table1.txt' overwrite into table table1; load data local inpath 'data_table2.txt' overwrite into table table2;

123456789101112131415161718192021222324252627

use test;DROP TABLE IF EXISTS table1;create table table1(student_no bigint comment '学号',student_name string comment '姓名')COMMENT 'test 学生信息'ROW FORMAT DELIMITEDFIELDS TERMINATED BY 't'LINES TERMINATED BY 'n'STORED AS TEXTFILE;DROP TABLE IF EXISTS table2;create table table2(student_no bigint comment '学号',class_no bigint comment '课程号')COMMENT 'test 学生选课信息'ROW FORMAT DELIMITEDFIELDS TERMINATED BY 't'LINES TERMINATED BY 'n'STORED AS TEXTFILE;load data local inpath 'data_table1.txt' overwrite into table table1;load data local inpath 'data_table2.txt' overwrite into table table2;

测试数据为:

hive left join测试数据

测试1:left join

语句:

select * from table1 left outer join table2 on(table1.student_no=table2.student_no);

结果:

FAILED: Parse Error: line 1:22 cannot recognize input near ‘left’ ‘join’ ‘table2’ in join type specifier

我用的HIVE版本是0.8,不支持直接的left join写法;

测试2:left outer join

语句:

select * from table1 left outer join table2 on(table1.student_no=table2.student_no);

结果:

1 name1 1 11 1 name1 1 12 1 name1 1 13 2 name2 2 11 2 name2 2 14 3 name3 3 15 3 name3 3 12 4 name4 4 13 4 name4 4 12 5 name5 5 14 5 name5 5 16 6 name6 NULL NULL

可以看到left outer join左边表的数据都列出来了,如果右边表没有对应的列,则写成了NULL值。

同时注意到,如果左边的主键在右边找到了N条,那么结果也是会叉乘得到N条的,比如这里主键为1的显示了右边的3条。

测试3:left semi join

语句:

select * from table1 left semi join table2 on(table1.student_no=table2.student_no);

结果:

1 name1 2 name2 3 name3 4 name4 5 name5

可以看到,只打印出了左边的表中的列,规律是如果主键在右边表中存在,则打印,否则过滤掉了。

结论:

hive不支持’left join’的写法; hive的left outer join:如果右边有多行和左边表对应,就每一行都映射输出;如果右边没有行与左边行对应,就输出左边行,右边表字段为NULL; hive的left semi join:相当于SQL的in语句,比如上面测试3的语句相当于“select * from table1 where table1.student_no in (table2.student_no)”,注意,结果中是没有B表的字段的。

0 人点赞