Question

我一直在努力使用弹簧批访问作业的作业参数。这是我到目前为止的实施。

@Configuration
@EnableBatchProcessing
@PropertySource("classpath:batch.properties")
public class CSVBatchServiceImpl extends StepExecutionListenerSupport implements CSVBatchService {
    private static final Logger LOGGER = LoggerFactory.getLogger(CSVBatchServiceImpl.class);
    @Autowired
    public JobBuilderFactory jobBuilderFactory;
    @Autowired
    public StepBuilderFactory stepBuilderFactory;

    private QuestionReader questionReader = new QuestionReader();

    @Bean(name = "importQuestionsJob")
    public Job importQuestionsJob() {
        return jobBuilderFactory.get("importQuestionsJob")
                .incrementer(new RunIdIncrementer())
                .flow(step1())
                .end()
                .build();
    }

    @Bean
    public Step step1() {
        return stepBuilderFactory.get("step1")
                .<Question, Question>chunk(2)
                .reader(questionReader.reader())
                .processor(processor())
                .build();
    }

    @Bean
    public QuestionProcessor processor() {
        return new QuestionProcessor();
    }
}

class QuestionReader extends StepExecutionListenerSupport {
    private static final Logger LOGGER = LoggerFactory.getLogger(QuestionReader.class);

    //TODO: remove this
    private static JsonNode getJsonNode(String str) {
        try {
            ObjectMapper mapper = new ObjectMapper();
            return mapper.readTree(str);
        } catch (IOException e) {
            throw new RuntimeException(e);
        }
    }

    @Bean
    public FlatFileItemReader<Question> reader() {
        FlatFileItemReader<Question> reader = new FlatFileItemReader<>();
        //TODO get this as a parameter
        reader.setResource(new ClassPathResource("duplicateLabels.csv"));
        reader.setLinesToSkip(1);
        reader.setLineMapper(new DefaultLineMapper<Question>() {{
            setLineTokenizer((new DelimitedLineTokenizer() {{
                setNames(new String[]{"label", "body", "real_answer"});
            }}));
            setFieldSetMapper(new QuestionFieldSetMapper());
        }});
        return reader;
    }

    private static class QuestionFieldSetMapper implements FieldSetMapper<Question> {
        public Question mapFieldSet(FieldSet fieldSet) {
            Question question = new Question();
            question.setLabel(fieldSet.readString(0));
            question.setBody(getJsonNode(fieldSet.readString(1)));
            question.setRealAnswer(getJsonNode(fieldSet.readString(2)));
            return question;
        }
    }
}

我称这项工作为：

JobParameters parameters = new JobParametersBuilder()
        .addLong("time", System.currentTimeMillis())
        .addString("filePath", "file.csv")
        .toJobParameters();
jobLauncher.run(importQuestionsJob, parameters);

如何访问reader函数中的filePath参数？

Answer 1

ItemProcessors批注是另一种对ItemReaders，ItemWriters，@BeforeStep等有效的解决方案。像StepExecutionListener一样由Eugene To mentioned支持。这是该解决方案的捷径。

一个实现可能看起来像这样

@BeforeStep
public void beforeStep(StepExecution stepExecution) {
    JobParameters jobParameters = stepExecution.getJobParameters();

    Long millis = jobParameters.getLong("time");
    String path = jobParameters.getString("filePath");
}

Answer 2

你应该可以做到，

@Value("#{jobParameters['filePath']}") String filePath;

如果出现任何问题，您可以尝试将您的阅读器放入@StepScope。

Answer 3

您可以在步骤中添加org.springframework.batch.core.listener.JobParameterExecutionContextCopyListener。

TaskletStep step = stepBuilderFactory.get("my-best-step") .<Item, Item>chunk(10) .reader(myBestReader) .writer(myBestWriter) .listener(new JobParameterExecutionContextCopyListener()) .build();

该监听器会将JobParameters复制到ItemReader的开放和更新方法中可用的ExecutionContext

Answer 4

访问作业参数的一种方法是对您的阅读器类实现StepExecutionListener，以利用其重写的beforeStep和afterStep方法，

@Override
public void beforeStep(StepExecution stepExecution) {
   String filePath = (String) stepExecution.getJobExecution().getExecutionContext()
        .get("filePath");
}

访问作业参数Spring Batch

4 个答案: