hive day05

排序

全局排序 order by

查询学生的成绩,按照分数降序排列

select * from student  s left join score sco on s.id=sco.id order by sco.s_score desc

查询学生的成绩，并按照分数升序排列

 select * from student s left join score sco on s.s_id = sco.s_id order sco.s_score asc;

desc规定排序为降序,默认order by排序为升序.

按照别名排序

按照分数的平均值排序

select s_id ,avg(s_score) avg from score group by s_id order by avg;

多个列排序

按照学生id和平均成绩进行排序

select s_id ,avg(s_score) avg from score group by s_id order by s_id,avg;

每个MapReduce内部排序（Sort By）局部排序
Sort By：每个MapReduce内部进行排序，对全局结果集来说不是排序。
①设置reduce个数

set mapreduce.job.reduces=3;

②查看设置reduce个数

set mapreduce.job.reduces;

③查询成绩按照成绩降序排列

select * from score sort by s_score;

将查询结果导入到文件中（按照成绩降序排列）

insert overwrite local directory '/export/servers/hivedatas/sort' select * from score sort by s_score;

分区排序

distribute by:类似于MapReduce中的partition,进行分区结合sort by使用.
注:在hive中要求distribute by语句要写在sort by语句之前.
例:先按照学生id进行分区,再按照学生成绩进行排序.
①设置reduce个数,将对应的id划分到对应的reduce中

hive>set mapreduce job.reduces=7;

②通过distribute by进行数据的分区

hive>insert overwrite local directory '/home/sort' select * from scoredistribute by s_id sort by s_score;

cluster by

当distributie by 和sort by的字段相同时,可以使用cluster by
cluster by除了具有distribute by 的功能之外还有sort by的功能,但是排序只能是倒序排序,不能指定desc

写法
hive>select * from score cluster by;
等价于
hive>select * from score distribute by s_id sort by s_id;