Spring Batch官网介绍:
A lightweight, comprehensive batch framework designed to enable the development of robust batch applications vital for the daily operations of enterprise systems.(一款轻量的、全面的批处理框架,用于开发强大的日常运营的企业级批处理应用程序。)
springbatch
主要实现批量数据的处理,我对batch进行的封装,提出了jobBase类型,具体job需要实现它即可。Spring Batch 不仅提供了统一的读写接口、丰富的任务处理方式、灵活的事务管理及并发处理,同时还支持日志、监控、任务重启与跳过等特性,大大简化了批处理应用开发,将开发人员从复杂的任务配置管理过程中解放出来,使他们可以更多地去关注核心的业务处理过程。
几个组件
•job
•step
•read
•write
•listener
•process
•validator
JobBase定义了几个公用的方法
/**
* springBatch的job基础类.
*/
public abstract class JobBase {
/**
* 批次.
*/
protected int chunkCount = 5000;
/**
* 监听器.
*/
private JobExecutionListener jobExecutionListener;
/**
* 处理器.
*/
private ValidatingItemProcessor validatingItemProcessor;
/**
* job名称.
*/
private String jobName;
/**
* 检验器.
*/
private Validator validator;
@Autowired
private JobBuilderFactory job;
@Autowired
private StepBuilderFactory step;
/**
* 初始化.
*
* @param jobName job名称
* @param jobExecutionListener 监听器
* @param validatingItemProcessor 处理器
* @param validator 检验
*/
public JobBase(String jobName,
JobExecutionListener jobExecutionListener,
ValidatingItemProcessor validatingItemProcessor,
Validator validator) {
this.jobName = jobName;
this.jobExecutionListener = jobExecutionListener;
this.validatingItemProcessor = validatingItemProcessor;
this.validator = validator;
}
/**
* job初始化与启动.
*/
public Job getJob() throws Exception {
return job.get(jobName).incrementer(new RunIdIncrementer())
.start(syncStep())
.listener(jobExecutionListener)
.build();
}
/**
* 执行步骤.
*
* @return
*/
public Step syncStep() throws Exception {
return step.get("step1")
.chunk(chunkCount)
.reader(reader())
.processor(processor())
.writer(writer())
.build();
}
/**
* 单条处理数据.
*
* @return
*/
public ItemProcessor processor() {
validatingItemProcessor.setValidator(processorValidator());
return validatingItemProcessor;
}
/**
* 校验数据.
*
* @return
*/
@Bean
public Validator processorValidator() {
return validator;
}
/**
* 批量读数据.
*
* @return
* @throws Exception
*/
public abstract ItemReader reader() throws Exception;
/**
* 批量写数据.
*
* @return
*/
@Bean
public abstract ItemWriter writer();
}
主要规定了公用方法的执行策略,而具体的job名称,读,写还是需要具体JOB去实现的。
具体Job实现
@Configuration
@EnableBatchProcessing
public class SyncPersonJob extends JobBase {
@Autowired
private DataSource dataSource;
@Autowired
@Qualifier("primaryJdbcTemplate")
private JdbcTemplate jdbcTemplate;
/**
* 初始化,规则了job名称和监视器.
*/
public SyncPersonJob() {
super("personJob", new PersonJobListener(), new PersonItemProcessor(), new BeanValidator<>());
}
@Override
public ItemReader reader() throws Exception {
StringBuffer sb = new StringBuffer();
sb.append("select * from person");
String sql = sb.toString();
JdbcCursorItemReader jdbcCursorItemReader =
new JdbcCursorItemReader<>();
jdbcCursorItemReader.setSql(sql);
jdbcCursorItemReader.setRowMapper(new BeanPropertyRowMapper<>(Person.class));
jdbcCursorItemReader.setDataSource(dataSource);
return jdbcCursorItemReader;
}
@Override
@Bean("personJobWriter")
public ItemWriter writer() {
JdbcBatchItemWriter writer = new JdbcBatchItemWriter();
writer.setItemSqlParameterSourceProvider(new BeanPropertyItemSqlParameterSourceProvider());
String sql = "insert into person_export " + "(id,name,age,nation,address) "
+ "values(:id, :name, :age, :nation,:address)";
writer.setSql(sql);
writer.setDataSource(dataSource);
return writer;
}
}
写操作需要定义自己的bean的声明
注意,需要为每个job的write启个名称,否则在多job时,write将会被打乱
/**
* 批量写数据.
*
* @return
*/
@Override
@Bean("personVerson2JobWriter")
public ItemWriter writer() {
}
添加一个api,手动触发
@Autowired
SyncPersonJob syncPersonJob;
@Autowired
JobLauncher jobLauncher;
void exec(Job job) throws Exception {
JobParameters jobParameters = new JobParametersBuilder()
.addLong("time", System.currentTimeMillis())
.toJobParameters();
jobLauncher.run(job, jobParameters);
}
@RequestMapping("/run1")
public String run1() throws Exception {
exec(syncPersonJob.getJob());
return "personJob success";
}
总结
以上所述是小编给大家介绍的springbatch的封装与使用实例详解,希望对大家有所帮助,如果大家有任何疑问欢迎给我留言,小编会及时回复大家的!