spring hadoop之batch处理(二)

一、测试
public class MrBatchApp {
    // Log
    private static final Log log = LogFactory.getLog(MrBatchApp.class);
    //
    public static void main(String[] args) throws      JobParametersInvalidException, JobExecutionAlreadyRunningException, JobRestartException, JobInstanceAlreadyCompleteException {
    System.out.println("TEST");
    // 加载对应的xml配置文件
    AbstractApplicationContext context = new ClassPathXmlApplicationContext("classpath:/META-INF/spring/*-context.xml");
        log.info("Batch Tweet Hashtag MR Job Running");
        // 关闭"钩子" 为了方便在适当的时候关闭 spring ioc
        // （在非web环境下，关闭spring ioc需要手动完成）
        context.registerShutdownHook();
        // job 发射器
        // JobLaucher是一个简化的job的控制接口；基于运行时不同的标识
        // 该接口并不能确保执行job是同步还是异步
        JobLauncher jobLauncher = context.getBean(JobLauncher.class);
        // job
        Job job = context.getBean(Job.class);
        // 运行job
        jobLauncher.run(job, new JobParameters());

    }
}
二、xml配置文件
（1）、common 配置
<?xml version="1.0" encoding="UTF-8"?>
<beans xmlns="http://www.springframework.org/schema/beans"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:p="http://www.springframework.org/schema/p"
xsi:schemaLocation="http://www.springframework.org/schema/beans http://www.springframework.org/schema/beans/spring-beans.xsd">
// job 仓库
<bean id="jobRepository" class="org.springframework.batch.core.repository.support.MapJobRepositoryFactoryBean"/>
// 事务
<bean id="transactionManager" class="org.springframework.batch.support.transaction.ResourcelessTransactionManager"/>
// job launcher
<bean id="jobLauncher" class="org.springframework.batch.core.launch.support.SimpleJobLauncher" p:jobRepository-ref="jobRepository"/>
</beans>
(2)、特殊配置
<beans xmlns="http://www.springframework.org/schema/beans"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:batch="http://www.springframework.org/schema/batch"
xmlns:hdp="http://www.springframework.org/schema/hadoop"
xmlns:context="http://www.springframework.org/schema/context"
xsi:schemaLocation="http://www.springframework.org/schema/beans http://www.springframework.org/schema/beans/spring-beans.xsd
      http://www.springframework.org/schema/batch http://www.springframework.org/schema/batch/spring-batch.xsd
      http://www.springframework.org/schema/context http://www.springframework.org/schema/context/spring-context.xsd
      http://www.springframework.org/schema/hadoop http://www.springframework.org/schema/hadoop/spring-hadoop.xsd">
    // 引入common 配置
    <import resource="batch-common-context.xml"/>
    // hdfs uri/分析目录/统计目录/分析源文件
    <context:property-placeholder location="hadoop.properties"/>

<context:component-scan base-package="org.springframework.samples.hadoop.mapreduce" />
// 指定hdfs
<hdp:configuration>
        fs.defaultFS=${hd.fs}
</hdp:configuration>
    // 设定hdp 脚本 用于创建localSourceFile、inputDir、outputDir
    <hdp:script id="setupScript" location="file-prep.groovy" run-at-startup="true">
        <hdp:property name="localSourceFile" value="${localSourceFile}"/>
        <hdp:property name="inputDir" value="${tweets.input.path}"/>
        <hdp:property name="outputDir" value="${tweets.output.path}"/>
    </hdp:script>
    // 指定mapreduce step to step执行
    <!-- required since Hadoop Job is a class not an interface and we need to use a Job with step scope to access #{jobParameters['...']} -->
    <bean class="org.springframework.batch.core.scope.StepScope">
        <property name="proxyTargetClass" value="true"/>
    </bean>
    // 设置job steps
    <job id="job" xmlns="http://www.springframework.org/schema/batch">
        <step id="hashtagcount" next="result-step">
            <tasklet ref="hashtagcount-tasklet" />
        </step>
        <step id="result-step">
            <tasklet ref="results"/>
        </step>
    </job>
   
    <hdp:job-tasklet id="hashtagcount-tasklet" job-ref="hashtagcountJob" scope="step"/>

    <hdp:job id="hashtagcountJob"
        input-path="${tweets.input.path}"
        output-path="${tweets.output.path}"
        mapper="org.springframework.samples.hadoop.mapreduce.HashtagCount$TokenizerMapper"
        reducer="org.springframework.samples.hadoop.mapreduce.HashtagCount$LongSumReducer"
        scope="step" />
    // 指定统计结果 输出
    <hdp:script-tasklet id="results" scope="step">
        <hdp:script location="classpath:results.groovy">
            <hdp:property name="outputDir" value="${tweets.output.path}"/>
        </hdp:script>
    </hdp:script-tasklet>

</beans>
三、groovy脚本
// 判断分析源文件所在的目录是否存在 不存在创建 并将源文件复制到指定目录下
// 同时修改该文件夹的权限
if (!fsh.test(inputDir)) {
   fsh.mkdir(inputDir);
   fsh.copyFromLocal(localSourceFile, inputDir);
   fsh.chmod(700, inputDir)
}
// 判断统计结果目录是否存在 存在则删除
if (fsh.test(outputDir)) {
   fsh.rmr(outputDir)
}
-----------------------------------------------------------------------
// 输出分析统计结果的内容
println "RESULTS from " + outputDir
old = new File('results.txt')
if( old.exists() ) {
    old.delete()
}
fsh.get(outputDir + '/part-r-*', 'results.txt');
String fileContents = new File('results.txt').text
println fileContents
以上即可完全通过xml完成mapreduce的batch处理