时间戳

首页 > 技术文章 > 第六章_查询 【排序原理 order by、sort by、distribute by、cluster by】

第六章_查询 【排序原理 order by、sort by、distribute by、cluster by】

bajiaotai 2022-01-29 11:41 原文

1. order by(全局排序)
    1. 全局排序, 无论设置多少个 reduce个数,只会产生一个reduce
    2. 对大规模的数据集,全局排序效率非常低
1. 设置分区个数为 3
set mapreduce.job.reduces=3;

2. 运行sql
select name
      ,num
from (
    select '张飞' as name,1 as num
    union all
    select '张飞' as name,2 as num
    union all
    select '张飞' as name,3 as num
    union all
    select '赵云' as name,1 as num
    union all
    select '赵云' as name,2 as num
    union all
    select '赵云' as name,3 as num
) as t1
order by num desc ;

-- 查看结果(无论 reduceNum 设置为多少,只会产生一个 reduce)
Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 1

name    num
赵云    3
张飞    3
赵云    2
张飞    2
赵云    1
张飞    1
View Code
2. sort by (分区内排序)
    1. 为每个 reducer 内产生的文件排序
    2. 不指定 distribute by 时, 分区 = hash(查询字段) % 指定的reduce个数




-- 使用步骤
 -- 1. 设置 reduce 个数
 set mapreduce.job.reduces=2;

select name
      ,num
from (
    select '张飞' as name,1 as num
    union all
    select '张飞' as name,2 as num
    union all
    select '张飞' as name,3 as num
    union all
    select '赵云' as name,1 as num
    union all
    select '赵云' as name,2 as num
    union all
    select '赵云' as name,3 as num
) as t1
sort by num desc ;

--结果
INFO  : Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 2
--产生2个 reduce, 且按照 行的所有字段进行分区(也就是按照 name分区)

name    num
赵云  3
赵云  2
赵云  1
张飞  3
张飞  2
张飞  1




3. distribute by(指定 分区字段)
    1. 分区编号 = hash(指定分区字段) % reduceNub






insert overwrite local directory '/root/sortbykey'
select name
      ,num
from (
    select '张飞' as name,1 as num
    union all
    select '张飞' as name,2 as num
    union all
    select '张飞' as name,3 as num
    union all
    select '赵云' as name,1 as num
    union all
    select '赵云' as name,2 as num
    union all
    select '赵云' as name,3 as num
) as t1
distribute by name
sort by num desc ;

Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 3


-rw-r--r-- 1 root root 27 1月  29 11:12 000000_1
-rw-r--r-- 1 root root  0 1月  29 11:12 000001_2 (空文件)
-rw-r--r-- 1 root root 27 1月  29 11:12 000002_0

赵云3
赵云2
赵云1
张飞3
张飞2
张飞1



View Code


4. cluster by(分区内 排序)
    1. 当 distribute by 和 sort by 字段相同时, 可以使用cluster by
       默认为 升序排序, 不能指定 排序规则 desc 、asc
    2. cluster by name = distribute by name sort by name asc ;






insert overwrite local directory '/root/clusterby2'
select name
      ,num
from (
    select '张飞' as name,1 as num
    union all
    select '张飞' as name,2 as num
    union all
    select '张飞' as name,3 as num
    union all
    select '赵云' as name,1 as num
    union all
    select '赵云' as name,2 as num
    union all
    select '赵云' as name,3 as num
) as t1
cluster by name ;

name    num
赵云    3
赵云    2
赵云    1
张飞    3
张飞    2
张飞    1



View Code


5. mr排序连接: https://www.cnblogs.com/bajiaotai/p/15734910.html







	

		


推荐阅读