自定义函数hello，并注册到hive源码中并重新编译

1 编写自己的udf方法hello

package cn.zhangjin.hive.udf;
 
 
import org.apache.hadoop.hive.ql.exec.Description;
import org.apache.hadoop.hive.ql.exec.UDF;
import org.apache.hadoop.io.Text;
 
 
/**
 * @author zj
 * @create 2019-02-22 17:51
 * 一个UDF: hello
 */
 
 
@Description(name = "sayhello",
        value = "_FUNC_(input_str) - returns Hello : input_str ",
        extended = "Example:\n "
                + "  > SELECT _FUNC_('wxk') FROM src LIMIT 1;\n"
                + "  'Hello : wxk'\n")
public class hello extends UDF {
    public Text evaluate(Text input) {
        return new Text("Hello: " + input);
    }
}　　

pom配置

<url>http://maven.apache.org</url>
 
<properties>
    <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
    <hadoop.version>2.6.0-cdh5.7.0</hadoop.version>
    <hive.version>1.1.0-cdh5.7.0</hive.version>
</properties>
 
<repositories>
    <repository>
        <id>cloudera</id>
        <url>https://repository.cloudera.com/artifactory/cloudera-repos</url>
    </repository>
 
</repositories>
 
<!-- 设定插件仓库 -->
<pluginRepositories>
 
    <pluginRepository>
        <id>jeesite-repos</id>
        <name>Jeesite Repository</name>
        <url>http://maven.aliyun.com/nexus/content/groups/public</url>
    </pluginRepository>
 
</pluginRepositories>
 
<dependencies>
 
    <dependency>
        <groupId>org.apache.hadoop</groupId>
        <artifactId>hadoop-common</artifactId>
        <version>${hadoop.version}</version>
    </dependency>
 
    <dependency>
        <groupId>org.apache.hive</groupId>
        <artifactId>hive-exec</artifactId>
        <version>${hive.version}</version>
    </dependency>
 
    <dependency>
        <groupId>org.apache.hive</groupId>
        <artifactId>hive-jdbc</artifactId>
        <version>${hive.version}</version>
    </dependency>
 
    <dependency>
        <groupId>junit</groupId>
        <artifactId>junit</artifactId>
        <version>4.10</version>
        <scope>test</scope>
    </dependency>
</dependencies>

2 下载hive源码

源码下载：http://archive.cloudera.com/cdh5/cdh/5/hive-1.1.0-cdh5.7.0-src.tar.gz

参见：FunctionRegistry

3 自己修改代码

　（1）修改udf函数，并放入源码中　

将hello.java  放入 hive-1.1.0-cdh5.7.0/ql/src/java/org/apache/hadoop/hive/ql/udf 文件夹中
vi hello.java 
将 package com.****.hello; 修改为 package org.apache.hadoop.hive.ql.udf;

（2)修改FunctionRegistry.java 文件

vi hive-1.1.0-cdh5.7.0/ql/src/java/org/apache/hadoop/hive/ql/exec/FunctionRegistry.java 
文件头部 一长串 import 下添加，因为我们要吧这个UDF添加进去。
import org.apache.hadoop.hive.ql.udf.hello;
  
文件头部 static 块中添加  system.registerUDF("hello", hello.class, false);
如下：
static {
    system.registerGenericUDF("concat", GenericUDFConcat.class);
    system.registerUDF("hello", hello.class, false);
    system.registerUDF("substr", UDFSubstr.class, false);