一.创建Hive工程
1.1 新建Java工程
我们这里为了简单就不用Maven了，而是直接使用lib包并把我们需要的jar包放到lib当中。如下所示。
（lib这个目录是手工创建的）
image.png
1.2 将Hive的lib下所有jar包放到lib目录下
下载地址:https://downloads.apache.org/hive/hive-2.3.9/
备注:需要选择对应的版本
image.png
我们需要把hive的lib下的所有的包放到lib下，hive的jar包所在的位置如下图所示:
截图未截全
image.png
1.3 导入hadoop-common jar包
下载地址:https://repo1.maven.org/maven2/org/apache/hadoop/hadoop-common/3.0.3/
备注:需要选择对应的版本
image.png
1.4 将jar包添加到工程
我们把jar拷到hive-udf工程下的lib包后记得选中所有的jar包，然后右键在菜单中把鼠标放到"Build path"上，这时会出现子菜单，然后点击"Add to Build path"添加到工程当中。
image.png
image.png
二.编写java代码并打包
代码:
package myUdf;
import org.apache.hadoop.hive.ql.exec.UDF;
import org.apache.hadoop.io.Text;
/**
* @author 只是甲
* @date 2021-07-26
* @remark
*/
public class getDegree extends UDF {
public String evaluate(int degreetype) {
/* 1 -- 小学
2 -- 初中
3 -- 职业高中
4 -- 中专
5 -- 高中
6 -- 大专
7 -- 本科
8 -- 硕士
9 -- 博士
*/
String result;
if (degreetype == 1) {
result = "小学";
} else if (degreetype == 2) {
result = "初中";
} else if (degreetype == 3) {
result = "职业高中";
} else if (degreetype == 4) {
result = "中专";
} else if (degreetype == 5) {
result = "高中";
} else if (degreetype == 6) {
result = "大专";
} else if (degreetype == 7) {
result = "本科";
} else if (degreetype == 8) {
result = "硕士";
} else if (degreetype == 9) {
result = "博士";
} else {
result = "N/A";
}
return result;
}
}
工程概要如下:
image.png
导出jar包：
选中Project HiveUDF，右键，然后选中Export
image.png
image.png
三. 注册UDF
3.1 上传上一步的jar文件到服务器
image.png
3.2 注册UDF
接着我们需要注册UDF，通知hive我现在注册了一个UDF，要调用这个UDF。注册UDF需要在hive视图下进行，因此我们首先启动hive的bin目录下启动hive。
代码:
-- 添加jar包
add jar /home/java/getDegree.jar;
-- 查看现有的jar包
list jars;
-- 创建临时函数
create temporary function GetDegree as 'myUdf.getDegree';
测试记录:
hive> add jar /home/java/getDegree.jar;
Added [/home/java/getDegree.jar] to class path
Added resources: [/home/java/getDegree.jar]
hive> drop function GetDegree;
hive>
>
> list jars;
/home/java/getDegree.jar
hive> create temporary function GetDegree as 'myUdf.getDegree';
OK
Time taken: 0.023 seconds
hive>
3.3 Hive UDF测试
数据准备:
degree_type
1 -- 小学
2 -- 初中
3 -- 职业高中
4 -- 中专
5 -- 高中
6 -- 大专
7 -- 本科
8 -- 硕士
9 -- 博士
create table user_info(id int,degree_type int);
insert into user_info values (1,3);
insert into user_info values (2,1);
insert into user_info values (3,6);
insert into user_info values (4,4);
insert into user_info values (5,5);
insert into user_info values (6,9);
insert into user_info values (7,8);
insert into user_info values (8,2);
insert into user_info values (9,7);
hive>
> select * from user_info;
OK
user_info.id user_info.degree_type
1 3
2 1
3 6
4 4
5 5
6 9
7 8
8 2
9 7
Time taken: 0.088 seconds, Fetched: 9 row(s)
测试Hive UDF:
hive>
>
> select id,GetDegree(degree_type) from user_info order by id;
Query ID = root_20210726144704_e3235df0-fa86-4eec-8fcd-b5a5e87234cf
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>
21/07/26 14:47:04 INFO client.ConfiguredRMFailoverProxyProvider: Failing over to rm69
Starting Job = job_1627269473896_0006, Tracking URL = http://hp3:8088/proxy/application_1627269473896_0006/
Kill Command = /opt/cloudera/parcels/CDH-6.3.1-1.cdh6.3.1.p0.1470567/lib/hadoop/bin/hadoop job -kill job_1627269473896_0006
Hadoop job information for Stage-1: number of mappers: 2; number of reducers: 1
2021-07-26 14:47:10,383 Stage-1 map = 0%, reduce = 0%
2021-07-26 14:47:16,584 Stage-1 map = 100%, reduce = 0%, Cumulative CPU 5.19 sec
2021-07-26 14:47:22,759 Stage-1 map = 100%, reduce = 100%, Cumulative CPU 6.9 sec
MapReduce Total cumulative CPU time: 6 seconds 900 msec
Ended Job = job_1627269473896_0006
MapReduce Jobs Launched:
Stage-Stage-1: Map: 2 Reduce: 1 Cumulative CPU: 6.9 sec HDFS Read: 12073 HDFS Write: 342 HDFS EC Read: 0 SUCCESS
Total MapReduce CPU Time Spent: 6 seconds 900 msec
OK
1 职业高中
2 小学
3 大专
4 中专
5 高中
6 博士
7 硕士
8 初中
9 本科
Time taken: 20.596 seconds, Fetched: 9 row(s)
hive>
3.4 创建永久性的函数
上面一个步骤我们创建的是临时的函数，退出当前session后，该函数不可用。
[root@hp1 java]# hive
WARNING: Use "yarn jar" to launch YARN applications.
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/opt/cloudera/parcels/CDH-6.3.1-1.cdh6.3.1.p0.1470567/jars/log4j-slf4j-impl-2.8.2.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/opt/cloudera/parcels/CDH-6.3.1-1.cdh6.3.1.p0.1470567/jars/slf4j-log4j12-1.7.25.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
SLF4J: Actual binding is of type [org.apache.logging.slf4j.Log4jLoggerFactory]
Logging initialized using configuration in jar:file:/opt/cloudera/parcels/CDH-6.3.1-1.cdh6.3.1.p0.1470567/jars/hive-common-2.1.1-cdh6.3.1.jar!/hive-log4j2.properties Async: false
WARNING: Hive CLI is deprecated and migration to Beeline is recommended.
hive>
> use test;
OK
Time taken: 1.154 seconds
hive>
> select id,GetDegree(degree_type) from user_info order by id;
FAILED: SemanticException [Error 10011]: Invalid function GetDegree
此时我们可以来创建永久性的函数
-- OS端操作
sudo -u hdfs hadoop fs -mkdir /user/hive/lib
sudo -u hdfs hadoop fs -put /home/java/getDegree.jar /user/hive/lib
-- Hive命令行操作
CREATE FUNCTION GetDegree AS 'myUdf.getDegree' USING JAR 'hdfs:///user/hive/lib/getDegree.jar';
select id,GetDegree(degree_type) from user_info order by id;
测试记录:
[root@hp1 ~]# sudo -u hdfs hadoop fs -mkdir /user/hive/lib
[root@hp1 ~]#
[root@hp1 ~]#
[root@hp1 ~]# sudo -u hdfs hadoop fs -put /home/java/getDegree.jar /user/hive/lib
hive>
>
> CREATE FUNCTION GetDegree AS 'myUdf.getDegree' USING JAR 'hdfs://hp1:9866/user/hive/lib/getDegree.jar';
End of File Exception between local host is: "hp1/10.31.1.123"; destination host is: "hp1":9866; : java.io.EOFException; For more details see: http://wiki.apache.org/hadoop/EOFException
Failed to register test.getdegree using class myUdf.getDegree
FAILED: Execution Error, return code 1 from org.apache.hadoop.hive.ql.exec.FunctionTask
hive>
> CREATE FUNCTION GetDegree AS 'myUdf.getDegree' USING JAR 'hdfs:///user/hive/lib/getDegree.jar';
Added [/tmp/4e1c389c-4053-4197-9364-93a01840309c_resources/getDegree.jar] to class path
Added resources: [hdfs:///user/hive/lib/getDegree.jar]
OK
Time taken: 0.203 seconds
hive> select id,GetDegree(degree_type) from user_info order by id;
Query ID = root_20210726152617_b3da9372-0724-4126-8d62-f58fa55afc42
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>
21/07/26 15:26:18 INFO client.ConfiguredRMFailoverProxyProvider: Failing over to rm69
Starting Job = job_1627269473896_0007, Tracking URL = http://hp3:8088/proxy/application_1627269473896_0007/
Kill Command = /opt/cloudera/parcels/CDH-6.3.1-1.cdh6.3.1.p0.1470567/lib/hadoop/bin/hadoop job -kill job_1627269473896_0007
Hadoop job information for Stage-1: number of mappers: 2; number of reducers: 1
2021-07-26 15:26:26,852 Stage-1 map = 0%, reduce = 0%
2021-07-26 15:26:33,129 Stage-1 map = 100%, reduce = 0%, Cumulative CPU 2.6 sec
2021-07-26 15:26:38,290 Stage-1 map = 100%, reduce = 100%, Cumulative CPU 7.12 sec
MapReduce Total cumulative CPU time: 7 seconds 120 msec
Ended Job = job_1627269473896_0007
MapReduce Jobs Launched:
Stage-Stage-1: Map: 2 Reduce: 1 Cumulative CPU: 7.12 sec HDFS Read: 11873 HDFS Write: 342 HDFS EC Read: 0 SUCCESS
Total MapReduce CPU Time Spent: 7 seconds 120 msec
OK
1 职业高中
2 小学
3 大专
4 中专
5 高中
6 博士
7 硕士
8 初中
9 本科
Time taken: 22.51 seconds, Fetched: 9 row(s)
hive>
参考:
