DQL-排序

order by

全排序

reduce数只有一，不管设置没设置。

select d.name,e.name from dept d left join emp e on d.id = e.deptId order by d.id;

sort by

分区排序,通常与distribute by 连用

查看reduce数

set mapreduce.job.reduces;

设置reduce数

set mapreduce.job.reduces=3;

查询

select * from dept sort by dept.deptno;

distribute by

设定以什么分区

同样的设置reduce数

set mapreduce.job.redueces=6;

distribute by + sort by;

select d.no,d.name,e.no,e.name from dept d left join emp e on d.eno = e.no distribute by dept.no sort by e.no;

cluster by

相当于使用同一标准的distribute by + sort by;
例如：

select d.*,e.* from dept d left join emp e on d.eno = e.no
cluster by d.no;

select d.*,e.* from dept d left join emp e on d.eno=e.no 
distribute by d.no sort by d.no;