Hive 基本语法操练（一）：表操作

Hive 和 Mysql 的表操作语句类似，如果熟悉 Mysql，学习Hive 的表操作就非常容易了，下面对 Hive 的表操作进行深入讲解。

**(1)先来创建一个表名为student的内部表**

hive> create table if not exists student (sno INT, sname STRING, age INT, sex STRING) row format delimited fields terminated by '\t' stored as textfile;

OK

Time taken: 0.985 seconds

建表规则如下：

CREATE [EXTERNAL] TABLE [IF NOT EXISTS] table_name

  [(col_name data_type [COMMENT col_comment], ...)]

  [COMMENT table_comment]

  [PARTITIONED BY (col_name data_type [COMMENT col_comment], ...)]

  [CLUSTERED BY (col_name, col_name, ...)

  [SORTED BY (col_name [ASC|DESC], ...)] INTO num_buckets BUCKETS]

  [ROW FORMAT row_format]

  [STORED AS file_format]

  [LOCATION hdfs_path]

•CREATE TABLE 创建一个指定名字的表。如果相同名字的表已经存在，则抛出异常；用户可以用 IF NOT EXIST 选项来忽略这个异常

•EXTERNAL 关键字可以让用户创建一个外部表，在建表的同时指定一个指向实际数据的路径（LOCATION）

•LIKE 允许用户复制现有的表结构，但是不复制数据

•COMMENT可以为表与字段增加描述

•ROW FORMAT DELIMITED [FIELDS TERMINATED BY char] [COLLECTION ITEMS TERMINATED BY char]

[MAP KEYS TERMINATED BY char] [LINES TERMINATED BY char]

| SERDE serde_name [WITH SERDEPROPERTIES (property_name=property_value, property_name=property_value, ...)]

用户在建表的时候可以自定义 SerDe 或者使用自带的 SerDe。如果没有指定 ROW FORMAT 或者 ROW FORMAT DELIMITED，将会使用自带的 SerDe。在建表的时候，用户还需要为表指定列，用户在指定表的列的同时也会指定自定义的 SerDe，Hive 通过 SerDe 确定表的具体的列的数据。

•STORED AS

SEQUENCEFILE

| TEXTFILE

| RCFILE

| INPUTFORMAT input_format_classname OUTPUTFORMAT output_format_classname

如果文件数据是纯文本，可以使用 STORED AS TEXTFILE。如果数据需要压缩，使用 STORED AS SEQUENCE 。

**（2）创建外部表**

```

hive> create external table if not exists student2 (sno INT, sname STRING, age INT, sex STRING) row format delimited fields terminated by '\t' stored as textfile location '/user/external';

OK

Time taken: 0.089 seconds

hive> show tables;

OK

student1

student2

Time taken: 0.06 seconds, Fetched: 12 row(s)

```

（3）删除表

首先创建一个表名为test1的表

```

hive> create table if not exists test1(id INT, name STRING);

OK

Time taken: 0.064 seconds

```

然后查看一下是否有test1表

```

hive> show tables;

OK

student

student2

test1

Time taken: 0.22 seconds, Fetched: 3 row(s)

```

用命令删test1表

```

hive> drop table test1;

OK

Time taken: 0.838 seconds

```

查看test1表是否删除

```

hive> show tables;

OK

student

student2

Time taken: 0.14 seconds, Fetched: 2 row(s)

```

（4）修改表的结构，比如为表增加字段

首先看一下student表的结构

```

hive> desc student;

OK

sno                 	int

sname               	string

age                 	int

sex                 	string

Time taken: 0.142 seconds, Fetched: 4 row(s)

```

为表student增加两个字段

```

hive> alter table student add columns (address STRING, grade STRING);

OK

Time taken: 0.138 seconds

```

再查看一下表的结构，看是否增加

```

hive> desc student;

OK

sno                 	int

sname               	string

age                 	int

sex                 	string

address             	string

grade               	string

Time taken: 0.145 seconds, Fetched: 6 row(s)

```

（5）修改表名student为student1

```

hive> alter table student rename to student1;

OK

Time taken: 0.15 seconds

```

查看一下

```

hive> show tables;

OK

student1

student2

Time taken: 0.028 seconds, Fetched: 2 row(s)

```

（6）创建和已知表相同结构的表

```

hive> create table copy_student1 like student1;

OK

Time taken: 0.092 seconds

```

查看一下

```

hive> show tables;

OK

copy_student1

student1

student2

Time taken: 0.03 seconds, Fetched: 3 row(s)

```

***2、加入导入数据的方法，（数据里可以包含重复记录），只有导入了数据，才能供后边的查询使用***

**（1）加载本地数据load**

首先看一下表的结构

```

hive> desc student1;

OK

sno                 	int

sname               	string

age                 	int

sex                 	string

address             	string

grade               	string

Time taken: 0.118 seconds, Fetched: 6 row(s)

```

创建/home/hadoop/data目录，并在该目录下创建student1.txt文件，添加如下内容

```

[hadoop@master ~]$ cd /home

[hadoop@master home]$ ll

total 4

drwx------. 28 hadoop hadoop 4096 May 17 18:42 hadoop

[hadoop@master home]$ cd hadoop/

[hadoop@master ~]$ ll

total 32

drwxr-xr-x. 2 hadoop hadoop 4096 Apr  3 18:12 Desktop

drwxr-xr-x. 2 hadoop hadoop 4096 Apr  3 18:12 Documents

drwxr-xr-x. 2 hadoop hadoop 4096 Apr  3 18:12 Downloads

drwxr-xr-x. 2 hadoop hadoop 4096 Apr  3 18:12 Music

drwxr-xr-x. 2 hadoop hadoop 4096 Apr  3 18:12 Pictures

drwxr-xr-x. 2 hadoop hadoop 4096 Apr  3 18:12 Public

drwxr-xr-x. 2 hadoop hadoop 4096 Apr  3 18:12 Templates

drwxr-xr-x. 2 hadoop hadoop 4096 Apr  3 18:12 Videos

[hadoop@master ~]$ sudo mkdir data/

[hadoop@master ~]$ cd data/

[hadoop@master data]$ sudo vim student1.txt

  201501001       张三    22      男      北京    大三

  201501003       李四    23      男      上海    大二

  201501004       王娟    22      女      广州    大三

  201501010       周王    24      男      深圳    大四

  201501011       李红    23      女      北京    大三

```

加载数据到student1表中

```

hive> load data local inpath '/home/hadoop/data/student1.txt' into table student1;

Loading data to table default.student1

Table default.student1 stats: [numFiles=1, numRows=0, totalSize=300, rawDataSize=0]

OK

Time taken: 1.271 seconds

```

查看是否加载成功

```

hive> select * from student1;

OK

201501001	张三	22	男	北京	大三

201501003	李四	23	男	上海	大二

201501004	王娟	22	女	广州	大三

201501010	周王	24	男	深圳	大四

201501011	李红	23	女	北京	大三

Time taken: 0.052 seconds, Fetched: 15 row(s)

```

**（2）加载hdfs中的文件**

首先将文件student1.txt上传到hdfs文件系统对应目录上

```

[hadoop@master hadoop-2.6.0]$ hadoop fs -put /home/hadoop/data/student1.txt /user/hive

[hadoop@master hadoop-2.6.0]$ hadoop fs -ls /user/hive

Found 2 items

-rw-r--r--   3 hadoop supergroup        193 2018-05-17 23:54 /user/hive/student1.txt

drwxr-xr-x   - hadoop supergroup          0 2018-05-17 23:10 /user/hive/warehouse

```

加载hdfs中的文件数据到copy_student1表中

```

hive> LOAD DATA INPATH '/user/hive/student1.txt' INTO TABLE copy_student1;

Loading data to table default.copy_student1

Table default.copy_student1 stats: [numFiles=1, totalSize=191]

OK

Time taken: 1.354 seconds

```

查看是否加载成功

```

hive> SELECT * FROM copy_student1;

OK

201501001	张三	22	男	北京	大三

201501003	李四	23	男	上海	大二

201501004	王娟	22	女	广州	大三

201501010	周王	24	男	深圳	大四

201501011	李红	23	女	北京	大三

Time taken: 0.44 seconds, Fetched: 5 row(s)

```

**（3）表插入数据（单表插入、多表插入）**

1）单表插入

首先创建一个表copy_student2,表结构和student1相同

```

hive> create table copy_student2 like student1;

OK

Time taken: 0.691 seconds

```

查看一下是否创建成功

```

hive> show tables;

OK

copy_student1

copy_student2

student1

student2

Time taken: 0.065 seconds, Fetched: 4 row(s)

```

看一下copy_student2表的表结构

```

hive> DESC copy_student2;

OK

sno                 	int

sname               	string

age                 	int

sex                 	string

address             	string

grade               	string

Time taken: 0.121 seconds, Fetched: 6 row(s)

```

把表student1中的数据插入到copy_student2表中

```

hive> insert overwrite table copy_student2 select * from copy_student1;

Query ID = hadoop_20180518000101_af36da39-e88b-4c1b-b89c-c000bf5f59dd

Total jobs = 3

Launching Job 1 out of 3

Number of reduce tasks is set to 0 since there's no reduce operator

Starting Job = job_1526553207632_0001, Tracking URL = http://master:8088/proxy/application_1526553207632_0001/

Kill Command = /opt/modules/hadoop-2.6.0/bin/hadoop job  -kill job_1526553207632_0001

Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 0

2018-05-18 00:01:16,715 Stage-1 map = 0%,  reduce = 0%

2018-05-18 00:01:29,632 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 1.36 sec

MapReduce Total cumulative CPU time: 1 seconds 360 msec

Ended Job = job_1526553207632_0001

Stage-4 is selected by condition resolver.

Stage-3 is filtered out by condition resolver.

Stage-5 is filtered out by condition resolver.

Moving data to: hdfs://ns/tmp/hive/hadoop/d6cb41c0-cc18-471e-861f-f08553caea48/hive_2018-05-18_00-01-00_086_4552315865937351442-1/-ext-10000

Loading data to table default.copy_student2

Table default.copy_student2 stats: [numFiles=1, numRows=5, totalSize=190, rawDataSize=185]

MapReduce Jobs Launched:

Stage-Stage-1: Map: 1   Cumulative CPU: 1.36 sec   HDFS Read: 403 HDFS Write: 268 SUCCESS

Total MapReduce CPU Time Spent: 1 seconds 360 msec

OK

Time taken: 35.015 seconds

```

查看数据是否插入

```

hive> select * from copy_student2;

OK

201501001	张三	22	男	北京	大三

201501003	李四	23	男	上海	大二

201501004	王娟	22	女	广州	大三

201501010	周王	24	男	深圳	大四

201501011	李红	23	女	北京	大三

Time taken: 0.073 seconds, Fetched: 5 row(s)

```

2）多表插入

先创建两个表

```

hive> CREATE TABLE copy_student3 LIKE student1;

OK

Time taken: 0.21 seconds

hive> CREATE TABLE copy_student4 LIKE student1;

OK

Time taken: 0.099 seconds

```

向多表插入数据

```

hive> FROM student1 INSERT OVERWRITE TABLE copy_student3 SELECT * INSERT OVERWRITE TABLE copy_student4 SELECT *;

(省略MapReduce过程)

```

查看结果

```

hive> select * from copy_student3;

OK

201501001	张三	22	男	北京	大三

201501003	李四	23	男	上海	大二

201501004	王娟	22	女	广州	大三

201501010	周王	24	男	深圳	大四

201501011	李红	23	女	北京	大三

Time taken: 0.049 seconds, Fetched: 5 row(s)

```

```

hive> select * from copy_student4;

OK

201501001	张三	22	男	北京	大三

201501003	李四	23	男	上海	大二

201501004	王娟	22	女	广州	大三

201501010	周王	24	男	深圳	大四

201501011	李红	23	女	北京	大三

Time taken: 0.049 seconds, Fetched: 5 row(s)

```

**3、有关表的内容的查询**

（1）查表的所有内容

```

hive> select * from student1;

OK

201501001	张三	22	男	北京	大三

201501003	李四	23	男	上海	大二

201501004	王娟	22	女	广州	大三

201501010	周王	24	男	深圳	大四

201501011	李红	23	女	北京	大三

Time taken: 0.041 seconds, Fetched: 5 row(s)

```

（2）查表的某个字段的属性

```

hive> select sname from student1;

OK

张三

李四

王娟

周王

李红

Time taken: 0.056 seconds, Fetched: 5 row(s)

```

（3）where条件查询

```

hive> SELECT * FROM student1 WHERE sno>201501004 AND address="北京";

OK

201501011	李红	23	女	北京	大三

Time taken: 0.203 seconds, Fetched: 1 row(s)

```

（4）all和distinct的区别（这就要求表中要有重复的记录，或者某个字段要有重复的数据）

```

hive> select all age,grade from student1;

OK

22	大三

23	大二

22	大三

24	大四

23	大三

Time taken: 0.054 seconds, Fetched: 5 row(s)

```

```

hive> select age,grade from student1;

OK

22	大三

23	大二

22	大三

24	大四

23	大三

Time taken: 0.053 seconds, Fetched: 5 row(s)

```

```

hive> select distinct age,grade from student1;

Query ID = hadoop_20180518001414_fe7461b7-7edd-4661-abc4-14859e3dba91

Total jobs = 1

Launching Job 1 out of 1

Number of reduce tasks not specified. Estimated from input data size: 1

In order to change the average load for a reducer (in bytes):

  set hive.exec.reducers.bytes.per.reducer=<number>

In order to limit the maximum number of reducers:

  set hive.exec.reducers.max=<number>

In order to set a constant number of reducers:

  set mapreduce.job.reduces=<number>

Starting Job = job_1526553207632_0004, Tracking URL = http://master:8088/proxy/application_1526553207632_0004/

Kill Command = /opt/modules/hadoop-2.6.0/bin/hadoop job  -kill job_1526553207632_0004

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

2018-05-18 00:14:10,913 Stage-1 map = 0%,  reduce = 0%

2018-05-18 00:14:22,260 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 1.27 sec

2018-05-18 00:14:36,734 Stage-1 map = 100%,  reduce = 100%, Cumulative CPU 2.51 sec

MapReduce Total cumulative CPU time: 2 seconds 510 msec

Ended Job = job_1526553207632_0004

MapReduce Jobs Launched:

Stage-Stage-1: Map: 1  Reduce: 1   Cumulative CPU: 2.51 sec   HDFS Read: 391 HDFS Write: 40 SUCCESS

Total MapReduce CPU Time Spent: 2 seconds 510 msec

OK

22	大三

23	大三

23	大二

24	大四

Time taken: 34.358 seconds, Fetched: 4 row(s)

```

```

hive> select distinct age from student1;

Query ID = hadoop_20180518001414_69278499-54b5-42b7-867c-4ebe8113a2f9

Total jobs = 1

Launching Job 1 out of 1

Number of reduce tasks not specified. Estimated from input data size: 1

In order to change the average load for a reducer (in bytes):

  set hive.exec.reducers.bytes.per.reducer=<number>

In order to limit the maximum number of reducers:

  set hive.exec.reducers.max=<number>

In order to set a constant number of reducers:

  set mapreduce.job.reduces=<number>

Starting Job = job_1526553207632_0005, Tracking URL = http://master:8088/proxy/application_1526553207632_0005/

Kill Command = /opt/modules/hadoop-2.6.0/bin/hadoop job  -kill job_1526553207632_0005

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

2018-05-18 00:14:56,548 Stage-1 map = 0%,  reduce = 0%

2018-05-18 00:15:03,047 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 0.85 sec

2018-05-18 00:15:10,390 Stage-1 map = 100%,  reduce = 100%, Cumulative CPU 1.98 sec

MapReduce Total cumulative CPU time: 1 seconds 980 msec

Ended Job = job_1526553207632_0005

MapReduce Jobs Launched:

Stage-Stage-1: Map: 1  Reduce: 1   Cumulative CPU: 1.98 sec   HDFS Read: 391 HDFS Write: 9 SUCCESS

Total MapReduce CPU Time Spent: 1 seconds 980 msec

OK

22

23

24

Time taken: 23.181 seconds, Fetched: 3 row(s)

```

（5）limit限制查询

```

hive> SELECT * FROM student1 LIMIT 4;

OK

201501001	张三	22	男	北京	大三

201501003	李四	23	男	上海	大二

201501004	王娟	22	女	广州	大三

201501010	周王	24	男	深圳	大四

Time taken: 0.253 seconds, Fetched: 4 row(s)

```

**(6) GROUP BY 分组查询**

group by 分组查询在数据统计时比较常用，接下来讲解 group by 的使用。

1) 创建一个表 group_test，表的内容如下。

```

create table group_test(uid STRING, gender STRING, ip STRING) row format delimited fields terminated by '\t' stored as textfile;

OK

Time taken: 0.449 seconds

```

```

[hadoop@master test]$ sudo vim user.txt

08	female  192.168.1.42

01	male    192.168.1.22

02	female  192.168.1.3

01	male    192.168.1.26

03	male    192.168.1.5

08	female  192.168.1.62

04	male    192.168.1.9

06	female  192.168.1.52

06	female  192.168.1.7

08	female  192.168.1.21

05	male    192.168.1.8

01	male	192.168.1.2

01	male	192.168.1.32

05	male    192.168.1.29

03	male    192.168.1.23

06	female  192.168.1.201

07	female  192.168.1.11

08	female  192.168.1.88

```

向 group_test 表中导入数据。

hive> load data local inpath '/home/hadoop/test/user.txt' into table group_test;

Loading data to table default.group_test

Table default.group_test stats: [numFiles=1, totalSize=193]

OK

Time taken: 0.865 seconds

2) 计算表的行数命令如下。

```

hive> select count(*) from group_test;

Query ID = hadoop_20180518040808_a73617a5-dd9a-48c4-b2a9-0ce1dd4bf4cd

Total jobs = 1

Launching Job 1 out of 1

Number of reduce tasks determined at compile time: 1

In order to change the average load for a reducer (in bytes):

  set hive.exec.reducers.bytes.per.reducer=<number>

In order to limit the maximum number of reducers:

  set hive.exec.reducers.max=<number>

In order to set a constant number of reducers:

  set mapreduce.job.reduces=<number>

Starting Job = job_1526553207632_0013, Tracking URL = http://master:8088/proxy/application_1526553207632_0013/

Kill Command = /opt/modules/hadoop-2.6.0/bin/hadoop job  -kill job_1526553207632_0013

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

2018-05-18 04:08:45,431 Stage-1 map = 0%,  reduce = 0%

2018-05-18 04:08:58,184 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 1.59 sec

2018-05-18 04:09:09,818 Stage-1 map = 100%,  reduce = 100%, Cumulative CPU 2.78 sec

MapReduce Total cumulative CPU time: 2 seconds 780 msec

Ended Job = job_1526553207632_0013

MapReduce Jobs Launched:

Stage-Stage-1: Map: 1  Reduce: 1   Cumulative CPU: 2.78 sec   HDFS Read: 624 HDFS Write: 3 SUCCESS

Total MapReduce CPU Time Spent: 2 seconds 780 msec

OK

18

Time taken: 34.896 seconds, Fetched: 1 row(s)hive> create table group_gender_sum(gender STRING, sum INT);

OK

Time taken: 0.081 seconds

```

3) 根据性别计算去重用户数。

首先创建一个表 group_gender_sum

```

hive> create table group_gender_sum(gender STRING,sum INT);

OK

Time taken: 0.142 seconds

```

将表 group_test 去重后的数据导入表 group_gender_sum。

```

hive> insert overwrite table group_gender_sum select group_test.gender,count(distinct group_test.uid) from group_test group by group_test.gender;

Query ID = hadoop_20180518041010_e51ae2fb-0b9e-4b5d-9a0c-87946496282f

Total jobs = 1

Launching Job 1 out of 1

Number of reduce tasks not specified. Estimated from input data size: 1

In order to change the average load for a reducer (in bytes):

  set hive.exec.reducers.bytes.per.reducer=<number>

In order to limit the maximum number of reducers:

  set hive.exec.reducers.max=<number>

In order to set a constant number of reducers:

  set mapreduce.job.reduces=<number>

Starting Job = job_1526553207632_0014, Tracking URL = http://master:8088/proxy/application_1526553207632_0014/

Kill Command = /opt/modules/hadoop-2.6.0/bin/hadoop job  -kill job_1526553207632_0014

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

2018-05-18 04:10:44,336 Stage-1 map = 0%,  reduce = 0%

2018-05-18 04:10:50,573 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 0.82 sec

2018-05-18 04:10:58,903 Stage-1 map = 100%,  reduce = 100%, Cumulative CPU 3.12 sec

MapReduce Total cumulative CPU time: 3 seconds 120 msec

Ended Job = job_1526553207632_0014

Loading data to table default.group_gender_sum

Table default.group_gender_sum stats: [numFiles=1, numRows=17, totalSize=371, rawDataSize=354]

MapReduce Jobs Launched:

Stage-Stage-1: Map: 1  Reduce: 1   Cumulative CPU: 3.12 sec   HDFS Read: 624 HDFS Write: 452 SUCCESS

Total MapReduce CPU Time Spent: 3 seconds 120 msec

OK

Time taken: 29.357 seconds

```

同时可以做多个聚合操作，但是不能有两个聚合操作有不同的 distinct 列。下面正确合法的聚合操作语句。

首先创建一个表 group_gender_agg

```

hive> create table group_gender_agg(gender STRING, sum1 INT, sum2 INT, sum3 INT);

OK

Time taken: 0.092 seconds

```

将表 group_test 聚合后的数据插入表 group_gender_agg。

```

hive> insert overwrite table group_gender_agg select group_test.gender,count(distinct group_test.uid),count(*),sum(distinct group_test.uid) from group_test group by group_test.gender;

Query ID = hadoop_20180518041212_0cf81102-2c8f-4370-8cda-3b7d61c51877

Total jobs = 1

Launching Job 1 out of 1

Number of reduce tasks not specified. Estimated from input data size: 1

In order to change the average load for a reducer (in bytes):

  set hive.exec.reducers.bytes.per.reducer=<number>

In order to limit the maximum number of reducers:

  set hive.exec.reducers.max=<number>

In order to set a constant number of reducers:

  set mapreduce.job.reduces=<number>

Starting Job = job_1526553207632_0015, Tracking URL = http://master:8088/proxy/application_1526553207632_0015/

Kill Command = /opt/modules/hadoop-2.6.0/bin/hadoop job  -kill job_1526553207632_0015

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

2018-05-18 04:12:45,953 Stage-1 map = 0%,  reduce = 0%

2018-05-18 04:12:52,218 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 0.81 sec

2018-05-18 04:12:59,519 Stage-1 map = 100%,  reduce = 100%, Cumulative CPU 2.42 sec

MapReduce Total cumulative CPU time: 2 seconds 420 msec

Ended Job = job_1526553207632_0015

Loading data to table default.group_gender_agg

Table default.group_gender_agg stats: [numFiles=1, numRows=17, totalSize=439, rawDataSize=422]

MapReduce Jobs Launched:

Stage-Stage-1: Map: 1  Reduce: 1   Cumulative CPU: 2.42 sec   HDFS Read: 624 HDFS Write: 520 SUCCESS

Total MapReduce CPU Time Spent: 2 seconds 420 msec

OK

Time taken: 21.103 seconds

```

但是，不允许在同一个查询内有多个 distinct 表达式。下面的查询是不允许的。

```

hive> insert overwrite table group_gender_agg select group_test.gender,count(distinct group_test.uid),count(distinct group_test.ip) from group_test group by group_test.gender;

```

这条查询语句是不合法的，因为 distinct group_test.uid 和 distinct group_test.ip 操作了uid 和 ip 两个不同的列。

(7) ORDER BY 排序查询

ORDER BY 会对输入做全局排序，因此只有一个 Reduce（多个 Reduce 无法保证全局有序）会导致当输入规模较大时，需要较长的计算时间。使用 ORDER BY 查询的时候，为了优化查询的速度，使用 hive.mapred.mode 属性。

```

hive.mapred.mode = nonstrict;(default value/默认值)

hive.mapred.mode=strict;

```

与数据库中 ORDER BY 的区别在于，在 hive.mapred.mode=strict 模式下必须指定limit ，否则执行会报错。

```

hive> set hive.mapred.mode=strict;

hive> select * from group_test order by uid limit 5;

Query ID = hadoop_20180518041414_f4daefe3-60ec-43d3-ab5c-d7fa7518fc5c

Total jobs = 1

Launching Job 1 out of 1

Number of reduce tasks determined at compile time: 1

In order to change the average load for a reducer (in bytes):

  set hive.exec.reducers.bytes.per.reducer=<number>

In order to limit the maximum number of reducers:

  set hive.exec.reducers.max=<number>

In order to set a constant number of reducers:

  set mapreduce.job.reduces=<number>

Starting Job = job_1526553207632_0016, Tracking URL = http://master:8088/proxy/application_1526553207632_0016/

Kill Command = /opt/modules/hadoop-2.6.0/bin/hadoop job  -kill job_1526553207632_0016

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

2018-05-18 04:14:18,047 Stage-1 map = 0%,  reduce = 0%

2018-05-18 04:14:25,572 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 0.98 sec

2018-05-18 04:14:31,896 Stage-1 map = 100%,  reduce = 100%, Cumulative CPU 2.07 sec

MapReduce Total cumulative CPU time: 2 seconds 70 msec

Ended Job = job_1526553207632_0016

MapReduce Jobs Launched:

Stage-Stage-1: Map: 1  Reduce: 1   Cumulative CPU: 2.07 sec   HDFS Read: 624 HDFS Write: 121 SUCCESS

Total MapReduce CPU Time Spent: 2 seconds 70 msec

OK

01	male	192.168.1.32

01	male	192.168.1.2

01	male    192.168.1.22

01	male    192.168.1.26

02	female  192.168.1.3

Time taken: 22.228 seconds, Fetched: 5 row(s)

```

(8) SORT BY 查询

sort by 不受 hive.mapred.mode 的值是否为 strict 和 nostrict 的影响。sort by 的数据只能保证在同一个 Reduce 中的数据可以按指定字段排序。

使用 sort by 可以指定执行的 Reduce 个数（set mapred.reduce.tasks=< number>）这样可以输出更多的数据。对输出的数据再执行归并排序，即可以得到全部结果。

```

hive> set hive.mapred.mode=strict;

hive> select * from group_test sort by uid ;

Query ID = hadoop_20180518041616_68543eaf-2bac-4c35-bad6-dd286052ded6

Total jobs = 1

Launching Job 1 out of 1

Number of reduce tasks not specified. Estimated from input data size: 1

In order to change the average load for a reducer (in bytes):

  set hive.exec.reducers.bytes.per.reducer=<number>

In order to limit the maximum number of reducers:

  set hive.exec.reducers.max=<number>

In order to set a constant number of reducers:

  set mapreduce.job.reduces=<number>

Starting Job = job_1526553207632_0017, Tracking URL = http://master:8088/proxy/application_1526553207632_0017/

Kill Command = /opt/modules/hadoop-2.6.0/bin/hadoop job  -kill job_1526553207632_0017

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

2018-05-18 04:16:11,201 Stage-1 map = 0%,  reduce = 0%

2018-05-18 04:16:19,537 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 0.83 sec

2018-05-18 04:16:26,844 Stage-1 map = 100%,  reduce = 100%, Cumulative CPU 1.88 sec

MapReduce Total cumulative CPU time: 1 seconds 880 msec

Ended Job = job_1526553207632_0017

MapReduce Jobs Launched:

Stage-Stage-1: Map: 1  Reduce: 1   Cumulative CPU: 1.88 sec   HDFS Read: 624 HDFS Write: 469 SUCCESS

Total MapReduce CPU Time Spent: 1 seconds 880 msec

OK

01	male	192.168.1.32

01	male	192.168.1.2

01	male    192.168.1.22

01	male    192.168.1.26

02	female  192.168.1.3

03	male    192.168.1.5

03	male    192.168.1.23

04	male    192.168.1.9

05	male    192.168.1.29

05	male    192.168.1.8

06	female  192.168.1.7

06	female  192.168.1.52

06	female  192.168.1.201

07	female  192.168.1.11

08	female  192.168.1.88

08	female  192.168.1.21

08	female  192.168.1.62

08	female  192.168.1.42

Time taken: 26.065 seconds, Fetched: 18 row(s)

```

(9) DISTRIBUTE BY 排序查询

按照指定的字段对数据划分到不同的输出 Reduce 文件中，操作如下。

```

hive> insert overwrite local directory '/home/hadoop/djt/test' select * from group_test distribute by length(gender);

```

此方法根据 gender 的长度划分到不同的 Reduce 中，最终输出到不同的文件中。length 是内建函数，也可以指定其它的函数或者使用自定义函数。

```

hive> insert overwrite local directory '/home/hadoop/djt/test' select * from group_test order by gender  distribute by length(gender);

```

order by gender 与 distribute by length(gender) 不能共用。

(10) CLUSTER BY 查询

cluster by 除了具有 distribute by 的功能外还兼具 sort by 的功能。

以上就是博主为大家介绍的这一板块的主要内容，这都是博主自己的学习过程，希望能给大家带来一定的指导作用，有用的还望大家点个支持，如果对你没用也望包涵，有错误烦请指出。如有期待可关注博主以第一时间获取更新哦，谢谢！　

 版权声明：本文为博主原创文章，未经博主允许不得转载。
巴特西

Hive 基本语法操练（一）：表操作

最新文章

热门文章
