代码之家 › 专栏 › 技术社区 › Josh

Spring批处理平面文件头

flat-file spring-batch spring

Josh · 技术社区 · 7 年前

我有个头像是这样的:

// Writer
    @Bean(name = "cms200Writer")
    @StepScope
    public FlatFileItemWriter<Cms200Item> cmsWriter(@Value("#{jobExecutionContext}") Map<Object, Object> ec, //
            @Qualifier("cms200LineAggregator") FormatterLineAggregator<Cms200Item> lineAgg) throws IOException {
        @SuppressWarnings("unchecked")
        String fileName = ((Map<String, MccFtpFile>) ec.get(AbstractSetupTasklet.BATCH_FTP_FILES)).get("cms").getLocalFile();
        //Ensure the file can exist.
        PrintWriter fos = getIoHarness().getFileOutputStream(fileName);
        fos.close();
        FlatFileItemWriter<Cms200Item> writer = new FlatFileItemWriter<>();
        writer.setResource(new FileSystemResource(fileName));
        writer.setLineAggregator(lineAgg);
        Calendar cal = Calendar.getInstance();
        Date date = cal.getTime();
        DateFormat dateFormat = new SimpleDateFormat("HH:mm:ss");
        String formattedDate=dateFormat.format(date);
        writer.setHeaderCallback(new FlatFileHeaderCallback() {
            public void writeHeader(Writer writer) throws IOException  {
                writer.write("                                            Test Company.                                                           " + formattedDate);
                writer.write("\n  CMS200                  CUSTOMER SHIPMENT MANIFEST AUTHORIZATION BY CUSTOMER NAME                                        Page 1");
                writer.write("\n\n");
                writer.write("      CUSTOMER  NAME          CITY        ST  CONTROL  MNFST ID  AUTH CODE   I03  CLS  EDI  EXPRESS   POV     MOST CURRENT  DEACTIVE");
                writer.write("\n");
                writer.write("                                                NBR                               TRL  214   WORK                ACCESS       DATE  ");
            }
        });
        return writer;
    }

每次处理53条记录时,我都要打印此标题。我不知道如何在我的Spring批处理工作中实现这个逻辑。我有 writeCount 添加到我的执行上下文中,但不确定如何访问它,或者这是否是正确的方法。

我发布的writer在我的BatchConfiguration.java文件中

编辑:

下面是我的文件步骤和添加的块大小

@Bean(name = "cms200FileStep")
    public Step createFileStep(StepBuilderFactory stepFactory, //
            @Qualifier("cms200Reader") ItemReader<Cms200Item> reader, //
            Cms200Processor processor, //
            @Qualifier("cms200Writer") ItemWriter<Cms200Item> writer) {
        return stepFactory.get("cms200FileStep") //
                .<Cms200Item, Cms200Item>chunk(100000) //
                .reader(reader) //
                .processor(processor) //
                .writer(writer).chunk(53) //
                .allowStartIfComplete(true)//
                .build();//
    }

编辑:添加的作业配置

// Job
    @Bean(name = "mccCMSCLRPTjob")
    public Job mccCmsclrptjob(JobBuilderFactory jobFactory, //
            @Qualifier("cms200SetupStep") Step setupStep, //
            @Qualifier("cms200FileStep") Step fileStep, //
            @Qualifier("putFtpFilesStep") Step putFtpStep, //
            @Qualifier("cms200TeardownStep") Step teardownStep, //
            @Autowired SingleInstanceListener listener,
            @Autowired ChunkSizeListener chunkListener) { //
        return jobFactory.get("mccCMSCLRPTjob") //
                .incrementer(new RunIdIncrementer()) //
                .listener(listener) //
                .start(setupStep) //
                .next(fileStep) //
                .next(putFtpStep) //
                .next(teardownStep) //
                .build();
    }

编辑: 添加侦听器

   @Bean(name = "cms200FileStep")
        public Step createFileStep(StepBuilderFactory stepFactory, //
                @Qualifier("cms200Reader") ItemReader<Cms200Item> reader, //
                Cms200Processor processor, //
                @Qualifier("cms200Writer") ItemWriter<Cms200Item> writer,
                @Autowired ChunkSizeListener listener) {
            return stepFactory.get("cms200FileStep") //
                    .<Cms200Item, Cms200Item>chunk(100000) //
                    .reader(reader) //
                    .processor(processor) //
                    .writer(writer).chunk(53) //
                    .allowStartIfComplete(true)//
                    .listener(listener) //
                    .build();//
        }

编辑: 经过多次反复,这就是我现在所处的位置

    // Utility Methods
    @Bean(name = "cms200FileStep")
    public Step createFileStep(StepBuilderFactory stepFactory, Map<Object, Object> ec, //
            @Qualifier("cms200Reader") ItemReader<Cms200Item> reader, //
            Cms200Processor processor, //
            @Qualifier("cms200Writer") ItemWriter<Cms200Item> writer) throws IOException {
        @SuppressWarnings("unchecked")
        String fileName = ((Map<String, MccFtpFile>) ec.get(AbstractSetupTasklet.BATCH_FTP_FILES)).get("cms").getLocalFile();
        return stepFactory.get("cms200FileStep") //
                .<Cms200Item, Cms200Item>chunk(100000) //
                .reader(reader) //
                .processor(processor) //
                .writer(writer).chunk(53) //
                .allowStartIfComplete(true)//
//                .listener((ChunkListener) listener) //
                .listener((ChunkListener) new ChunkSizeListener(new File(fileName))) //
                .build();//
    }

0 回复 | 直到 7 年前

Mahmoud Ben Hassine 7 年前

这个 FlatFileHeaderCallback 只在面向块的步骤之前调用一次,也就是在所有块之前调用。

每次处理53条记录时,我都要打印此标题

您可以将块大小设置为53并使用 ChunkListener 或 ItemWriteListener 写入所需的数据。

编辑:添加示例

class MyChunkListener extends StepListenerSupport {

    private FileWriter fileWriter;

    public MyChunkListener(File file) throws IOException {
        this.fileWriter = new FileWriter(file, true);
    }

    @Override
    public void beforeChunk(ChunkContext context) {
        try {
            fileWriter.write("your custom header");
            fileWriter.flush();
        } catch (IOException e) {
            System.err.println("Unable to write header to file");
        }
    }

    @Override
    public ExitStatus afterStep(StepExecution stepExecution) {
        try {
            fileWriter.close();
        } catch (IOException e) {
            System.err.println("Unable to close writer");
        }
        return super.afterStep(stepExecution);
    }
}