Hive学习笔记(一)各种连接

表和数据

• 创建数据库

create database if not exists test;
use test;

• 创建sales表

create table sales(
		name  varchar(50),
		product_id	int
		)
row format delimited fields terminated by ',';

• 创建things表

create table things(
		produt_id int,
		product_name varchar(100)
		)
row format delimited fields terminated by ',';

• 装载数据
  默认在/home/usename下找

load data local inpath 'hive/data/sales.txt'
overwrite into table sales;

load data local inpath 'hive/data/things.txt'
overwrite into table things;

• sales.txt
  wei,1
  duan,2
  tao,3
  yu,0

• things.txt
  1,laptop
  2,cup
  3,car
  4,bar

笛卡尔积 join

select *
from sales s join things t;

inner join

select * from sales s join things t on s.product_id==t.product_id;

left outer join

select * from sales s left outer join things t on s.product_id==t.product_id;

right outer join

select * from sales s right outer join things t on s.product_id==t.product_id;

full outer join

select * from sales s full outer join things t on s.product_id==t.product_id;

left semi join

select * from sales s left semi join things t on s.product_id==t.product_id;

right semi join

无
select * from sales s right semi join things t on s.product_id==t.product_id;

map-side join

适用于大小表　小表放在map端
select /+ mapjoin(t)/ * from sales s join things t on s.product_id==t.product_id;