Spring Batch 批处理(3) - Job、Flow、Split

Job的创建和调用

在成功创建一个job后,Spring Batch 默认在项目启动时候执行配置的job。往往在正常业务处理中,需要我们手动或者定时去触发job,所以这边便引入了jobLauncher、jobOperator两个执行器。

JobLauncher配置

此处我们通过web的API接口去调用 jobLauncher,通过接口传入job的参数。调用的Job 是根据 在创建job时候,Bean name去指定。

@Bean
public JobLauncher jobLauncher() {
    SimpleJobLauncher jobLauncher = new SimpleJobLauncher();
    jobLauncher.setJobRepository(jobRepository());
    jobLauncher.setTaskExecutor(new SimpleAsyncTaskExecutor()); //转换为异步任务
    jobLauncher.afterPropertiesSet();
    return jobLauncher;
}
   @Autowired
    private JobLauncher jobLauncher;
 
    @Autowired
    private Job jobLaunchDemoJob;
 
    @GetMapping("/{job1param}")
    public String runJob1(@PathVariable("job1param") String job1param) throws JobParametersInvalidException, JobExecutionAlreadyRunningException, JobRestartException, JobInstanceAlreadyCompleteException {
        System.out.println("Request to run job1 with param: " + job1param);
        JobParameters jobParameters = new JobParametersBuilder()
                .addString("job1param",job1param)
                .toJobParameters();
        jobLauncher.run(jobLaunchDemoJob,jobParameters);
        return "Job1 success.";
 
    }

JobOperator配置

    @Bean
    public JobOperator jobOperator(){
        SimpleJobOperator operator = new SimpleJobOperator();
 
        operator.setJobLauncher(jobLauncher);
        operator.setJobParametersConverter(new DefaultJobParametersConverter());
        operator.setJobRepository(jobRepository);
        operator.setJobExplorer(jobExplorer);
        operator.setJobRegistry(jobRegistry);
        return operator;
    }
    @Autowired
    private JobOperator jobOperator;
 
    @GetMapping("/{job2param}")
    public String runJob1(@PathVariable("job2param") String job2param) throws JobParametersInvalidException, JobExecutionAlreadyRunningException, JobRestartException, JobInstanceAlreadyCompleteException, JobInstanceAlreadyExistsException, NoSuchJobException {
        System.out.println("Request to run job2 with param: " + job2param);
 
        jobOperator.start("jobOperatorDemoJob","job2param="+job2param);
 
        return "Job2 success.";
 
    }

最后,定时任务调用,通过corn表达式,满足条件时候,即执行

    @Scheduled(fixedDelay = 5000)
    public void scheduler() throws JobInstanceAlreadyCompleteException, JobExecutionAlreadyRunningException, JobParametersInvalidException, JobRestartException, JobParametersNotFoundException, NoSuchJobException {
        jobOperator().startNextInstance("jobScheduledDemoJob");
    }




Job的嵌套


某些情况下,一个job需要嵌套执行另一个job,此时需要借助step。在一个step中嵌套执行一个子job,然后在父job中执行这个嵌套的step即可。

1.job可以嵌套使用,嵌套的Job我们将其称为子job,被嵌套的Job我们将其称为父job;

2.一个父Job可以有多个子Job;

3.子job不能单独运行,需要其父Job去启动;

如图所示

file




执行上图作业的程序配置如下:

    /*得到一个step*/
    @Bean
    public Step childStep(){
        return stepBuilderFactory.get("child").tasklet((StepContribution contribution, ChunkContext chunkContext) -> {
            System.out.println("子step执行……");
            return RepeatStatus.FINISHED;
        }).build();
    }

    /*得到一个job*/
    @Bean
    public Job childJob(){
        return jobBuilderFactory.get("childJob")
                .start(childStep())
                .build();
    }

    /*得到一个step,用于嵌套job*/
    /*
     * new StepBuilder("childStepJob")创建一个Step的工厂类,childStepJob指定一个Step的名字;
     * new JobStepBuilder创建一个工厂类,用来在step中嵌入一个job
     * repository用于存储job、step的执行信息,如果一个step已经执行完,一个job再次调用该step时,会提示step已经执行完毕,不会再执行
     * 该段代码意思:创建一个名字为childStepJob的step,在该step中嵌入执行一个名字为childJob的job
     * */
    @Bean
    public Step nestedStep(){
        return new JobStepBuilder(new StepBuilder("childStepJob"))
                .job(childJob())
                .repository(jobRepository)
                .build();
    }

    /*得到一个step,step的名字是parent*/
    @Bean
    public Step parentStep(){
        return stepBuilderFactory.get("parent").tasklet((StepContribution contribution, ChunkContext chunkContext) -> {
            System.out.println("父step执行……");
            return RepeatStatus.FINISHED;
        }).build();
    }

    /*得到一个job,job的名字为parent,注意不是parentJob,parentJob只是函数名*/
    @Bean
    public Job parentJob(){
        return jobBuilderFactory.get("parent")  
                .start(nestedStep())
                .next(parentStep())
                .build();
    }

}

在application.propertis中加入配置: ``` spring.batch.job.names=parentJob ```




程序中定义了childJob和parentJob两个job,childJob中执行的是名为child的step;parentJob中执行的名为parent和childStepJob的step,而childStepJob中执行的是名为childJob的job。

日志打印如下。

Job: [SimpleJob: [name=parentJob]] launched with the following parameters: [{}]
Executing step: [childStepJob]
Job: [SimpleJob: [name=childJob]] launched with the following parameters: [{}]
Executing step: [child]
子step执行……
Job: [SimpleJob: [name=childJob]] completed with the following parameters: [{}] and the following status: [COMPLETED]
Executing step: [parent]
父step执行……
Job: [SimpleJob: [name=parentJob]] completed with the following parameters: [{}] and the following status: [COMPLETED]

通过日志可以看出,只执行了parentJob,首先执行了parentJob中的名为childStepJob的step,执行childStepJob时执行了内嵌的名为childJob的Job;然后又执行了parentJob中名为parent的step




Job参数


1.JobParameters作用:在Job运行过程中,可以用来传递信息

2.通过”key---->value”键值对的形式传入,在代码中我们通过get(“key”)来获取value值

3.job的参数是在整个job的step的生命周期中都可以使用到,我们可以根据不同业务处理逻辑,传入所需参数。

调用过程,demo如下:

@Bean
    public Job myJobParametersDemoJob(){
        return jobBuilderFactory.get("myJobParametersDemoJob")
                .start(myJobParametersDemoStep())
                .build();
    }
 
    @Bean
    public Step myJobParametersDemoStep() {
        return stepBuilderFactory.get("myJobParametersDemoStep")
                .listener(this)
                .tasklet(((contribution, chunkContext) -> {
                    System.out.println("Parameter is : " + params.get("info"));
                    return RepeatStatus.FINISHED;
                })).build();
 
    }
 
    @Override
    public void beforeStep(StepExecution stepExecution) {
        params =  stepExecution.getJobParameters().getParameters();
 
    }
 
    @Override
    public ExitStatus afterStep(StepExecution stepExecution) {
        return null;
    }




Flow的创建和使用

Step 是一个独立的、顺序的处理步骤,包含完整的输入、处理以及输出。但是在企业应用中,我们面对的更多情况是多个步骤按照一定的顺序进行处理。因此如何维护步骤之间的执行顺序是我们需要考虑的。Spring Batch 提供了 Step Flow 来解决这个问题。

Flow有以下特点:

  1. flow是一个Step的集合,他规定了Step与Step之间的转换关系;

  2. 创建Flow可以达到复用的效果,让其在不同的Job之间进行复用;

  3. 使用FlowBuilder去创建一个Flow,他和Job类似,使用start(),next()以及end()来运行flow;


顺序Flow


``` // 创建Flow对象,指明Flow对象包含哪些Step @Bean public Flow jobFlowDemoFlow1(){ return new FlowBuilder("jobFlowDemoFlow1") .start(jobFlowDemoTwoStep1()) .next(jobFlowDemoTwoStep2()) .build(); }

@Bean
public Job jobFlowDemo(){
return jobBuilderFactory.get(“jobFlowDemo”)
.start(jobFlowDemoFlow1()).end()
.build();
}



 
 <br/> <br/>
## 决策器
 <br/>
动态自定义Decider 决定Flow 执行顺序

1.Decision:为我们提供下一步执行哪一个Step提供条件决策

2.JobExecutionDecider:接口,提供决策条件
 
 <br/>

1、自定义一个MyDecider, 根据调用次数,基数返回“EVEN”,偶数返回“ODD”;

public class MyDecider implements JobExecutionDecider {
private int count = 0;
@Override
public FlowExecutionStatus decide(JobExecution jobExecution, StepExecution stepExecution) {
count ++;
if (count % 2 == 0)
return new FlowExecutionStatus(“EVEN”);
else
return new FlowExecutionStatus(“ODD”);
}




 
 <br/>
 2、job中调用MyDecider,当返回“EVEN”时,调用evenStep;当返回“ODD”时,调用oddStep。

@Bean
public Step oddStep(){
    return stepBuilderFactory.get("oddStep")
            .tasklet(((stepContribution, chunkContext) -> {
                System.out.println("oddStep");
                return RepeatStatus.FINISHED;
            })).build();
}

@Bean
public Step evenStep(){
    return stepBuilderFactory.get("evenStep")
            .tasklet(((stepContribution, chunkContext) -> {
                System.out.println("evenStep");
                return RepeatStatus.FINISHED;
            })).build();
}

@Bean
public JobExecutionDecider myDecider(){
    return new MyDecider();
}

@Bean
public Job flowDecisonDemoJob(){
    return jobBuilderFactory.get("flowDecisonDemoJob").start(firstStep())
            .next(myDecider())
            .from(myDecider()).on("EVEN").to(evenStep())
            .from(myDecider()).on("ODD").to(oddStep())
            .from(oddStep()).on("*").to(myDecider())
            .end()
            .build();
}


<br/><br/><br/>


# 并行处理Split

<br/>

当遇到数据量比较大或者每个ITEM处理耗时的时候,能够并行处理是很重要的。并行处理的时候一般经历Split-Partitioned-Merge的过程

![file](https://graph.baidu.com/resource/2223b0a952f98c501fd7701583220953.png)

<br/>

// tag::jobstep[]
@Bean
public Job readandwritedbJob() {
Flow flow1 = new FlowBuilder(“asyncFlow1”).start(step2()).build();
Flow flow2 = new FlowBuilder(“asyncFlow2”).start(step3()).build();
return jobBuilderFactory.get(“readandwritedbJob”)
.incrementer(new RunIdIncrementer())
.flow(step1())
.split(new SimpleAsyncTaskExecutor()).add(flow1,flow2)
.end()
.build();
}

@Bean
public Tasklet tasklet1() {
    return new ReadAndWriteDbTask();
}

@Bean
public Tasklet tasklet2() {
    return new ReadAndWriteDbTask2();
}


@Bean
public Tasklet tasklet3() {
    return new ReadAndWriteDbTask3();
}

@Bean
public Step step1() {
    return stepBuilderFactory.get("step1")
            .tasklet(tasklet1())
            .build();
}

@Bean
public Step step2() {
    return stepBuilderFactory.get("step2")
            .tasklet(tasklet2())
            .build();
}

@Bean
public Step step3() {
    return stepBuilderFactory.get("step3")
            .tasklet(tasklet3())
            .build();
}
// end::jobstep[]





<br/><br/><br/>






# 监听Job Execution

<br/>

1.Listener:控制Job执行的一种方式

2.可以通过接口或者注解实现监听器

3.在spring-batch中提供各个级别的监听器接口,从job级别到item级别都有

(1)JobExecutionListener(before..,after..);

(2)StepExecutionListener(before..,after..);

(3)ChunkListener(before..,after..);

(4)ItemReaderListener;ItemWriterListener;ItemProcessListener(before..,after..,error..);


<br/>

当任务执行完毕或开始执行时,需要执行一些处理工作。这个时候可以使用JobExecutionListener:

public interface JobExecutionListener {
void beforeJob(JobExecution jobExecution);
void afterJob(JobExecution jobExecution);
}


<br/>
添加方式:

@Bean
public Job footballJob() {
return this.jobBuilderFactory.get(“footballJob”)
.listener(sampleListener()) //JobExecutionListener的实现类

.build();
}


<br/>
需要注意的是afterJob方法无论批处理任务成功还是失败都会被执行,所以增加以下判断:

public void afterJob(JobExecution jobExecution){
if( jobExecution.getStatus() == BatchStatus.COMPLETED ){
//job success
}
else if(jobExecution.getStatus() == BatchStatus.FAILED){
//job failure
}
}

<br/>
除了直接实现接口还可以用 @BeforeJob 和 @AfterJob 注解。



参考:

https://blog.csdn.net/wuzhiwei549/category_8573774.html

https://blog.51cto.com/13501268/2177746
发布了187 篇原创文章 · 获赞 79 · 访问量 16万+

猜你喜欢

转载自blog.csdn.net/weixin_38004638/article/details/104764978