Hive之Order,Sort,Cluster and Distribute By
- 测试数据
create table sort_test(
id int,
name string
)
row format delimited
fields terminated by '\t'
lines terminated by '\n'
stored as textfile;
[root@wadeyu ~]# cat sort_test.log
4679 aaa
4728 aaa
3040 aaa
4207 aaa
2231 aaa
1279 aaa
7954 aaa
582 aaa
7096 aaa
4878 aaa
9684 aaa
1540 aaa
4826 aaa
2543 aaa
2323 aaa
1420 aaa
5083 aaa
8965 aaa
1391 aaa
9719 aaa
9901 aaa
2393 aaa
6024 aaa
444 aaa
1574 aaa
8881 aaa
5739 aaa
8689 aaa
1614 aaa
9340 aaa
6726 aaa
109 aaa
6941 aaa
9562 aaa
9019 aaa
4945 aaa
2206 aaa
5910 aaa
8552 aaa
1795 aaa
2720 aaa
9007 aaa
8377 aaa
2179 aaa
3683 aaa
5869 aaa
5448 aaa
5223 aaa
5127 aaa
4616 aaa
2340 aaa
1268 aaa
4332 aaa
2989 aaa
19 aaa
7880 aaa
505 aaa
5975 aaa
5288 aaa
5682 aaa
376 aaa
7502 aaa
6448 aaa
3774 aaa
5541 aaa
9636 aaa
2037 aaa
246 aaa
6151 aaa
7837 aaa
1506 aaa
3749 aaa
9335 aaa
3973 aaa
5160 aaa
7929 aaa
834 aaa
3451 aaa
1766 aaa
6228 aaa
8961 aaa
8177 aaa
2340 aaa
4245 aaa
3226 aaa
2670 aaa
784 aaa
7699 aaa
2054 aaa
6006 aaa
4204 aaa
8905 aaa
6182 aaa
1271 aaa
5415 aaa
5164 aaa
4320 aaa
3736 aaa
2287 aaa
6559 aaa
- Order By
- Job中只会启动一个reduce做全局排序，数据量大时，耗时会很久
- 在strict模式(hive.mapred.mode=strict)下，必须添加limit语句限制返回条数
# 语法格式
colOrder: ( ASC | DESC )
colNullOrder: (NULLS FIRST | NULLS LAST) -- (Note: Available in Hive 2.1.0 and later)
orderBy: ORDER BY colName colOrder? colNullOrder? (',' colName colOrder? colNullOrder?)*
query: SELECT expression (',' expression)* FROM src orderBy
# 排序
select * from sort_test order by id desc;
+---------------+-----------------+--+
| sort_test.id | sort_test.name |
+---------------+-----------------+--+
| 9901 | aaa |
| 9719 | aaa |
| 9684 | aaa |
| 9636 | aaa |
| 9562 | aaa |
| 9340 | aaa |
| 9335 | aaa |
| 9019 | aaa |
| 9007 | aaa |
| 8965 | aaa |
| 8961 | aaa |
| 8905 | aaa |
| 8881 | aaa |
| 8689 | aaa |
| 8552 | aaa |
| 8377 | aaa |
| 8177 | aaa |
| 7954 | aaa |
| 7929 | aaa |
| 7880 | aaa |
| 7837 | aaa |
| 7699 | aaa |
| 7502 | aaa |
| 7096 | aaa |
| 6941 | aaa |
| 6726 | aaa |
| 6559 | aaa |
| 6448 | aaa |
| 6228 | aaa |
| 6182 | aaa |
| 6151 | aaa |
| 6024 | aaa |
| 6006 | aaa |
| 5975 | aaa |
| 5910 | aaa |
| 5869 | aaa |
| 5739 | aaa |
| 5682 | aaa |
| 5541 | aaa |
| 5448 | aaa |
| 5415 | aaa |
| 5288 | aaa |
| 5223 | aaa |
| 5164 | aaa |
| 5160 | aaa |
| 5127 | aaa |
| 5083 | aaa |
| 4945 | aaa |
| 4878 | aaa |
| 4826 | aaa |
| 4728 | aaa |
| 4679 | aaa |
| 4616 | aaa |
| 4332 | aaa |
| 4320 | aaa |
| 4245 | aaa |
| 4207 | aaa |
| 4204 | aaa |
| 3973 | aaa |
| 3774 | aaa |
| 3749 | aaa |
| 3736 | aaa |
| 3683 | aaa |
| 3451 | aaa |
| 3226 | aaa |
| 3040 | aaa |
| 2989 | aaa |
| 2720 | aaa |
| 2670 | aaa |
| 2543 | aaa |
| 2393 | aaa |
| 2340 | aaa |
| 2340 | aaa |
| 2323 | aaa |
| 2287 | aaa |
| 2231 | aaa |
| 2206 | aaa |
| 2179 | aaa |
| 2054 | aaa |
| 2037 | aaa |
| 1795 | aaa |
| 1766 | aaa |
| 1614 | aaa |
| 1574 | aaa |
| 1540 | aaa |
| 1506 | aaa |
| 1420 | aaa |
| 1391 | aaa |
| 1279 | aaa |
| 1271 | aaa |
| 1268 | aaa |
| 834 | aaa |
| 784 | aaa |
| 582 | aaa |
| 505 | aaa |
| 444 | aaa |
| 376 | aaa |
| 246 | aaa |
| 109 | aaa |
| 19 | aaa |
+---------------+-----------------+--+
- Sort By
- 排序前会根据排序字段分区，一个job启动多个reduce进行局部排序
- 如果有limit语句，会再次启动一个job，取出每个局部排好序的前n条，再进行全局排序
- 只保证局部有序，不保证全局有序
# Sort By语法
colOrder: ( ASC | DESC )
sortBy: SORT BY colName colOrder? (',' colName colOrder?)*
query: SELECT expression (',' expression)* FROM src sortBy
# 设置开启的reduce个数
set mapreduce.job.reduces=2;
0: jdbc:hive2://> set mapreduce.job.reduces;
+--------------------------+--+
| set |
+--------------------------+--+
| mapreduce.job.reduces=2 |
+--------------------------+--+
# 执行局部排序（未带limit）
0: jdbc:hive2://> select * from sort_test sort by id desc;
+---------------+-----------------+--+
| sort_test.id | sort_test.name |
+---------------+-----------------+--+
| 9901 | aaa |
| 9684 | aaa |
| 9340 | aaa |
| 9019 | aaa |
| 9007 | aaa |
| 8965 | aaa |
| 8961 | aaa |
| 8689 | aaa |
| 8552 | aaa |
| 8177 | aaa |
| 7837 | aaa |
| 7699 | aaa |
| 7502 | aaa |
| 6559 | aaa |
| 6448 | aaa |
| 6228 | aaa |
| 6024 | aaa |
| 6006 | aaa |
| 5975 | aaa |
| 5910 | aaa |
| 5869 | aaa |
| 5739 | aaa |
| 5682 | aaa |
| 5541 | aaa |
| 5448 | aaa |
| 5415 | aaa |
| 5288 | aaa |
| 5164 | aaa |
| 5160 | aaa |
| 5083 | aaa |
| 4878 | aaa |
| 4826 | aaa |
| 4679 | aaa |
| 4616 | aaa |
| 4245 | aaa |
| 4207 | aaa |
| 3736 | aaa |
| 3451 | aaa |
| 3226 | aaa |
| 3040 | aaa |
| 2989 | aaa |
| 2720 | aaa |
| 2670 | aaa |
| 2340 | aaa |
| 2231 | aaa |
| 2206 | aaa |
| 2054 | aaa |
| 2037 | aaa |
| 1766 | aaa |
| 1614 | aaa |
| 1540 | aaa |
| 1506 | aaa |
| 1420 | aaa |
| 1268 | aaa |
| 834 | aaa |
| 784 | aaa |
| 582 | aaa |
| 444 | aaa |
| 376 | aaa |
| 246 | aaa |
| 19 | aaa |
| 9719 | aaa |
| 9636 | aaa |
| 9562 | aaa |
| 9335 | aaa |
| 8905 | aaa |
| 8881 | aaa |
| 8377 | aaa |
| 7954 | aaa |
| 7929 | aaa |
| 7880 | aaa |
| 7096 | aaa |
| 6941 | aaa |
| 6726 | aaa |
| 6182 | aaa |
| 6151 | aaa |
| 5223 | aaa |
| 5127 | aaa |
| 4945 | aaa |
| 4728 | aaa |
| 4332 | aaa |
| 4320 | aaa |
| 4204 | aaa |
| 3973 | aaa |
| 3774 | aaa |
| 3749 | aaa |
| 3683 | aaa |
| 2543 | aaa |
| 2393 | aaa |
| 2340 | aaa |
| 2323 | aaa |
| 2287 | aaa |
| 2179 | aaa |
| 1795 | aaa |
| 1574 | aaa |
| 1391 | aaa |
| 1279 | aaa |
| 1271 | aaa |
| 505 | aaa |
| 109 | aaa |
+---------------+-----------------+--+
# 带limit排序(会额外再启动一个job进行全局排序)
0: jdbc:hive2://> select * from sort_test sort by id desc limit 300;
+---------------+-----------------+--+
| sort_test.id | sort_test.name |
+---------------+-----------------+--+
| 9901 | aaa |
| 9719 | aaa |
| 9684 | aaa |
| 9636 | aaa |
| 9562 | aaa |
| 9340 | aaa |
| 9335 | aaa |
| 9019 | aaa |
| 9007 | aaa |
| 8965 | aaa |
| 8961 | aaa |
| 8905 | aaa |
| 8881 | aaa |
| 8689 | aaa |
| 8552 | aaa |
| 8377 | aaa |
| 8177 | aaa |
| 7954 | aaa |
| 7929 | aaa |
| 7880 | aaa |
| 7837 | aaa |
| 7699 | aaa |
| 7502 | aaa |
| 7096 | aaa |
| 6941 | aaa |
| 6726 | aaa |
| 6559 | aaa |
| 6448 | aaa |
| 6228 | aaa |
| 6182 | aaa |
| 6151 | aaa |
| 6024 | aaa |
| 6006 | aaa |
| 5975 | aaa |
| 5910 | aaa |
| 5869 | aaa |
| 5739 | aaa |
| 5682 | aaa |
| 5541 | aaa |
| 5448 | aaa |
| 5415 | aaa |
| 5288 | aaa |
| 5223 | aaa |
| 5164 | aaa |
| 5160 | aaa |
| 5127 | aaa |
| 5083 | aaa |
| 4945 | aaa |
| 4878 | aaa |
| 4826 | aaa |
| 4728 | aaa |
| 4679 | aaa |
| 4616 | aaa |
| 4332 | aaa |
| 4320 | aaa |
| 4245 | aaa |
| 4207 | aaa |
| 4204 | aaa |
| 3973 | aaa |
| 3774 | aaa |
| 3749 | aaa |
| 3736 | aaa |
| 3683 | aaa |
| 3451 | aaa |
| 3226 | aaa |
| 3040 | aaa |
| 2989 | aaa |
| 2720 | aaa |
| 2670 | aaa |
| 2543 | aaa |
| 2393 | aaa |
| 2340 | aaa |
| 2340 | aaa |
| 2323 | aaa |
| 2287 | aaa |
| 2231 | aaa |
| 2206 | aaa |
| 2179 | aaa |
| 2054 | aaa |
| 2037 | aaa |
| 1795 | aaa |
| 1766 | aaa |
| 1614 | aaa |
| 1574 | aaa |
| 1540 | aaa |
| 1506 | aaa |
| 1420 | aaa |
| 1391 | aaa |
| 1279 | aaa |
| 1271 | aaa |
| 1268 | aaa |
| 834 | aaa |
| 784 | aaa |
| 582 | aaa |
| 505 | aaa |
| 444 | aaa |
| 376 | aaa |
| 246 | aaa |
| 109 | aaa |
| 19 | aaa |
+---------------+-----------------+--+
- Order By 和 Sort By区别
- Order By全局排序，Sort By局部排序
- 取TopN时，Sort By 比 Order By效率更高
- Distribute By
- 查询语句对指定字段分组
- 通常结合Sort By语句使用，比如同一个地区，不同商家排序，就需要用到这个
- Cluster By
- 分组且排序，等价于 Distribute By 和 Sort By 的结合
-- 使用示例
SELECT col1, col2 FROM t1 CLUSTER BY col1
SELECT col1, col2 FROM t1 DISTRIBUTE BY col1
SELECT col1, col2 FROM t1 DISTRIBUTE BY col1 SORT BY col1 ASC, col2 DESC
参考资料
【0】Hive wiki - LanguageManual SortBy
Hive之Order,Sort,Cluster and Distribute By的更多相关文章
- hive 中的Sort By、 Order By、Cluster By、Distribute By 区别
Order by: order by 会对输入做全局排序,因此只有一个reducer(多个reducer无法保证全局有序)只有一个reducer,会导致当输入规模较大时,需要较长的计算时间.在hive ...
- [转]hive中order by,distribute by,sort by,cluster by
转至http://my.oschina.net/repine/blog/296562 order by,distribute by,sort by,cluster by 查询使用说明 1 2 3 4 ...
- hive中order by、distribute by、sort by和cluster by的区别和联系
hive中order by.distribute by.sort by和cluster by的区别和联系 order by order by 会对数据进行全局排序,和oracle和mysql等数据库中 ...
- hive中order by ,sort by ,distribute by, cluster by 的区别（**很详细**）
hive 查询语法 select [all | distinct] select_ condition, select_ condition from table_name a [join table ...
- HiveQL之Sort by、Distribute by、Cluster by、Order By详解
在这里解释一下select语法中的order by.sort by.distribute by.cluster by.order by语法. 一.order by语法 在hiveQL中Order by ...
- hive中order by,sort by, distribute by, cluster by作用以及用法
1. order by Hive中的order by跟传统的sql语言中的order by作用是一样的,会对查询的结果做一次全局排序,所以说,只有hive的sql中制定了order by所有的 ...
- Hive中的order by、sort by、distribute by、cluster by解释及测试
结论: order by:全局排序,这也是4种排序手段中唯一一个能在终端输出中看出全局排序的方法,只有一个reduce,可能造成renduce任务时间过长,在严格模式下,要求必须具备limit子句. ...
- hive 排序 order by sort by distribute by cluster by
order by: order by是全局排序,受hive.mapred.mode的影响. 使用orderby有一些限制: 1.在严格模式下(hive.mapred.mod ...
- Hive中order by，sort by，distribute by，cluster by的区别
一:order by order by会对输入做全局排序,因此只有一个Reducer(多个Reducer无法保证全局有序),然而只有一个Reducer,会导致当输入规模较大时,消耗较长的计算时间.关于 ...
随机推荐
- java-IO操作性能对比
在软件系统中,IO速度比内存速度慢,IO读写在很多情况下会是系统的瓶颈. 在java标准IO操作中,InputStream和OutputStream提供基于流的IO操作,以字节为处理单位:Reader ...
- 洛谷 P2801 教主的魔法
题目描述 教主最近学会了一种神奇的魔法,能够使人长高.于是他准备演示给XMYZ信息组每个英雄看.于是N个英雄们又一次聚集在了一起,这次他们排成了一列,被编号为1.2.…….N. 每个人的身高一开始都是 ...
- swift 与 @objc
Objective-C entry points https://github.com/apple/swift-evolution/blob/master/proposals/0160-objc-in ...
- WebService 生成客户端
webservice 生成客户端 wsdl2java -encoding UTF-8 -p com.trm -d D:\happywork -client http://localhost/trm-t ...
- CAD参数绘制角度标注（网页版）
主要用到函数说明: _DMxDrawX::DrawDimAngular 绘制一个角度标注.详细说明如下: 参数 说明 DOUBLE dAngleVertexX 角度标注的顶点的X值 DOUBLE dA ...
- MFC如何设置窗口最前
首先,放到最前 this->SetWindowPos(&wndTopMost,0,0,0,0,SWP_NOMOVE|SWP_NOSIZE);//使窗口总是在最前面 this->Se ...
- java关于时间的相关操作
/** * 获取当天时间零点 * @return */ public Date gettoday(){ SimpleDateFormat sdf = new SimpleDateFormat(&quo ...
- 关于js事件冒泡和事件捕获
事件捕获指的是从document到触发事件的那个节点,即自上而下的去触发事件.相反的,事件冒泡是自下而上的去触发事件.绑定事件方法的第三个参数,就是控制事件触发顺序是否为事件捕获.true,事件捕获: ...
- 初识Python（windows）——下载、安装、使用
Table of Contents 1. Why is Python 1.1. Python和R 2. python的下载与安装 2.1. python的版本选择 2.2. python的下载 2.3 ...
- c++基础_字符串对比
#include <iostream> #include <string.h> #include <algorithm> using namespace std; ...