Spring Batch_異步併發的processor && writer

Spring Batch_異步併發的processor && writer

普通的配置一個job，在這個demo中：http://my.oschina.net/xinxingegeya/blog/343190

job的reader是經過遊標讀取，commit-interval="2"表示每讀取兩條數據，就要進行process，process完成以後就要進行write，process和write是同步進行的，也就是說

必須process兩條以後才能進行write，這二者不能異步進行。無疑，當process過程處理時間過長時，會拖慢整個過程的效率。還有process過程是single thread進行處理的，一個線程中處理兩條數據

比用兩個線程處理兩條數據效率要慢的多（當處理一條數據花費的時間比較多時），這樣會拖慢process過程的效率。

那麼如何提升整個批處理過程的效率？

1. 對於proceess和write過程異步化

2. 在process過程使用多線程處理數據

主要的代碼和配置：

spring-batch-async.xml

<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:context="http://www.springframework.org/schema/context"
	xsi:schemaLocation="http://www.springframework.org/schema/beans http://www.springframework.org/schema/beans/spring-beans-4.0.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">

	<!-- 包的掃描 -->
	<context:component-scan base-package="com.lyx.batch" />

	<bean id="exceptionHandler" class="com.lyx.batch.ExceptionListener" />

	<batch:step id="abstractStep" abstract="true">
		<batch:listeners>
			<batch:listener ref="exceptionHandler" />
		</batch:listeners>
	</batch:step>
	<bean id="abstractCursorReader" abstract="true"
		class="org.springframework.batch.item.database.JdbcCursorItemReader">
		<property name="dataSource" ref="dataSource" />
	</bean>

	<batch:job id="addPeopleDescJob">
		<batch:step id="addDescStep" parent="abstractStep">
			<batch:tasklet>
				<batch:chunk reader="peopleAddDescReader" processor="asyncProcessor"
					writer="asynWriter" commit-interval="2" />
			</batch:tasklet>
		</batch:step>
	</batch:job>
	<bean id="peopleAddDescReader" parent="abstractCursorReader"
		scope="step">
		<property name="sql">
			<value><![CDATA[select first_name ,last_name from people where 
			first_name like ? or last_name like ?]]></value>
		</property>
		<property name="rowMapper" ref="peopleRowMapper" />
		<property name="preparedStatementSetter" ref="preparedStatementSetter" />
		<property name="fetchSize" value="20" />
	</bean>

	<!--配置異步併發的processor和wirter -->
	<bean id="asyncProcessor" class="com.lyx.batch.AsyncPeopleAddDescItemProcessor">
		<property name="taskExecutor" ref="asyncExecutor" />
	</bean>
	<bean id="asyncExecutor" class="org.springframework.core.task.SimpleAsyncTaskExecutor" />
	<bean id="asynWriter" class="com.lyx.batch.AsyncPeopleAddDescItemWriter" />
	<bean id="peopleRowMapper" class="com.lyx.batch.PeopleRowMapper" />
	<bean id="preparedStatementSetter" class="com.lyx.batch.PeoplePreparedStatementSetter" />

	<!--tomcat jdbc pool數據源配置 -->
	<bean id="dataSource" class="org.apache.tomcat.jdbc.pool.DataSource"
		destroy-method="close">
		<property name="poolProperties">
			<bean class="org.apache.tomcat.jdbc.pool.PoolProperties">
				<property name="driverClassName" value="com.mysql.jdbc.Driver" />
				<property name="url" value="jdbc:mysql://localhost:3306/test" />
				<property name="username" value="root" />
				<property name="password" value="034039" />
			</bean>
		</property>
	</bean>

	<!-- spring batch 配置jobRepository -->
	<batch:job-repository id="jobRepository"
		data-source="dataSource" transaction-manager="transactionManager"
		isolation-level-for-create="REPEATABLE_READ" table-prefix="BATCH_"
		max-varchar-length="1000" />
	<!-- spring的事務管理器 -->
	<bean id="transactionManager"
		class="org.springframework.jdbc.datasource.DataSourceTransactionManager">
		<property name="dataSource" ref="dataSource" />
	</bean>

	<!-- batch luncher -->
	<bean id="jobLauncher"
		class="org.springframework.batch.core.launch.support.SimpleJobLauncher">
		<property name="jobRepository" ref="jobRepository" />
	</bean>
</beans>

AsyncPeopleAddDescItemProcessor.java

package com.lyx.batch;

import java.util.concurrent.Callable;
import java.util.concurrent.Future;
import java.util.concurrent.FutureTask;

import org.springframework.batch.item.ItemProcessor;
import org.springframework.core.task.SyncTaskExecutor;
import org.springframework.core.task.TaskExecutor;

public class AsyncPeopleAddDescItemProcessor implements
		ItemProcessor<People, Future<PeopleDESC>> {

	/**
	 * SyncTaskExecutor會在當前線程執行完客戶提交給它的任務，即它是以同步方式完成任務的執行的。
	 * 
	 */
	private TaskExecutor taskExecutor = new SyncTaskExecutor();

	public Future<PeopleDESC> process(final People item) throws Exception {
		FutureTask<PeopleDESC> task = new FutureTask<PeopleDESC>(
				new Callable<PeopleDESC>() {
					public PeopleDESC call() throws Exception {
						Thread.sleep(2000);
						System.out.println("process people desc");
						return new PeopleDESC(item.getLastName(),
								item.getFirstName(), Thread.currentThread()
										.getName());
					}
				});
		this.taskExecutor.execute(task);
		return task;
	}

	public TaskExecutor getTaskExecutor() {
		return this.taskExecutor;
	}

	public void setTaskExecutor(TaskExecutor taskExecutor) {
		this.taskExecutor = taskExecutor;
	}

}

AsyncPeopleAddDescItemWriter.java

package com.lyx.batch;

import java.util.LinkedList;
import java.util.List;
import java.util.concurrent.Future;

import javax.sql.DataSource;

import org.springframework.batch.item.ItemWriter;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.jdbc.core.JdbcTemplate;
import org.springframework.stereotype.Component;

@Component
public class AsyncPeopleAddDescItemWriter implements
		ItemWriter<Future<PeopleDESC>> {

	private JdbcTemplate jdbcTemplate;

	@Autowired
	public void setDataSource(DataSource dataSource) {
		this.jdbcTemplate = new JdbcTemplate(dataSource);
	}

	public void write(List<? extends Future<PeopleDESC>> items)
			throws Exception {
		LinkedList<Future<PeopleDESC>> linklist = new LinkedList<Future<PeopleDESC>>(
				items);

		Future<PeopleDESC> future;

		// the head of this linklist
		while ((future = linklist.poll()) != null) {
			if (future.isDone()) {
				if (!future.isCancelled()) {
					System.out.println("write people desc");
					PeopleDESC peopleDESC = future.get();
					this.jdbcTemplate
							.update("insert into ok_people (first_name, last_name, batch_desc) values (?, ?, ?)",
									peopleDESC.getFirstName(),
									peopleDESC.getLastName(),
									peopleDESC.getDesc());
				}
			} else {
				linklist.addLast(future);
			}
		}
	}
}

 

AppMain5.java

package com.lyx.batch;

import org.springframework.batch.core.ExitStatus;
import org.springframework.batch.core.Job;
import org.springframework.batch.core.JobExecution;
import org.springframework.batch.core.JobParametersBuilder;
import org.springframework.batch.core.JobParametersInvalidException;
import org.springframework.batch.core.launch.JobLauncher;
import org.springframework.batch.core.repository.JobExecutionAlreadyRunningException;
import org.springframework.batch.core.repository.JobInstanceAlreadyCompleteException;
import org.springframework.batch.core.repository.JobRestartException;
import org.springframework.context.ApplicationContext;
import org.springframework.context.support.ClassPathXmlApplicationContext;

/**
 * spring batch 的異步任務執行 reader processor writer之間的異步執行 程序運行時間： 108884ms
 * 
 * @author Lenovo
 *
 */
public class AppMain5 {

	public static void main(String[] args)
			throws JobExecutionAlreadyRunningException, JobRestartException,
			JobInstanceAlreadyCompleteException, JobParametersInvalidException {

		long startTime = System.currentTimeMillis(); // 獲取開始時間

		@SuppressWarnings("resource")
		ApplicationContext context = new ClassPathXmlApplicationContext(
				new String[] { "classpath:spring-batch-async.xml" });
		JobParametersBuilder jobParametersBuilder = new JobParametersBuilder();
		Job job = (Job) context.getBean("addPeopleDescJob");
		JobLauncher launcher = (JobLauncher) context.getBean("jobLauncher");
		JobExecution result = launcher.run(job,
				jobParametersBuilder.toJobParameters());
		ExitStatus es = result.getExitStatus();
		if (es.getExitCode().equals(ExitStatus.COMPLETED.getExitCode())) {
			System.out.println("任務正常完成");
		} else {
			System.out.println("任務失敗，exitCode=" + es.getExitCode());
		}

		long endTime = System.currentTimeMillis(); // 獲取結束時間
		System.out.println("程序運行時間： " + (endTime - startTime) + "ms");
	}

}

以上就是主要的關於異步併發的代碼。

若是使用異步併發的效果是很是明顯的，前提條件是處理一條數據花費時間過長，若是不是很長，在極短的時間內就能處理完，使用異步併發和不使用異步併發差異不是很明顯。

下面是我沒有使用異步併發作測試的類，在process過程當中thread.sleep()，2秒鐘的時間，同時，異步併發的process也是thread.sleep()兩秒的時間。運行一下看看時間上的差異：

（處理的數據是100條，前者是沒有使用異步併發，後者使用異步併發）

 * 程序運行時間： 213657ms 程序運行時間： 108884ms

AddPeopleDescProcessor.java

package com.lyx.batch;

import org.springframework.batch.item.ItemProcessor;

public class AddPeopleDescProcessor implements
		ItemProcessor<People, PeopleDESC> {

	public PeopleDESC process(People item) throws Exception {
		Thread.sleep(2000);
		System.out.println("process people desc");
		return new PeopleDESC(item.getLastName(), item.getFirstName(), Thread
				.currentThread().getName());
	}

}

AddDescPeopleWriter.java

package com.lyx.batch;

import java.util.List;

import javax.sql.DataSource;

import org.springframework.batch.item.ItemWriter;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.jdbc.core.JdbcTemplate;
import org.springframework.stereotype.Component;

@Component
public class AddDescPeopleWriter implements ItemWriter<PeopleDESC> {

	private JdbcTemplate jdbcTemplate;

	@Autowired
	public void setDataSource(DataSource dataSource) {
		this.jdbcTemplate = new JdbcTemplate(dataSource);
	}

	public void write(List<? extends PeopleDESC> items) throws Exception {
		for (PeopleDESC peopleDESC : items) {
			System.out.println("write people desc");
			this.jdbcTemplate
					.update("insert into ok_people (first_name, last_name, batch_desc) values (?, ?, ?)",
							peopleDESC.getFirstName(),
							peopleDESC.getLastName(), peopleDESC.getDesc());
		}
	}

}

=============END=============