如何在 spring 批处理中的单个项目阅读器中读取逗号分隔和 pipe 行分隔的 csv 文件

Question

I am new to sprig batch. I have a folder which contain multiple csv file, I have implemented MultiResourceItemReader () to read those file. It is working only if all csv file are pipe line ("|") separated.

I want to read both comma (",") separated csv and pipe line separated csv using single reader. Is it possible? if yes how?

Here is my code

@Bean
@StepScope
public MultiResourceItemReader<Person> multiResourceItemReader(@Value("#{jobParameters[x]}") String x,@Value("#{jobParameters[y]}") String y,@Value("#{jobParameters[z]}") String z) {

    Resource[] resourcessss = null;
    ClassLoader cl = this.getClass().getClassLoader();
    ResourcePatternResolver resolver = new PathMatchingResourcePatternResolver(cl);
    try {
        resourcessss = resolver.getResources("file:" + z);
    }catch(Exception e) {

    }
    MultiResourceItemReader<Person> resourceItemReader = new MultiResourceItemReader<Person>();
    resourceItemReader.setResources(resourcessss);
    resourceItemReader.setDelegate(reader());

    return resourceItemReader;
}



@Bean
public FlatFileItemReader<Person> reader() {
    FlatFileItemReader<Person> reader = new FlatFileItemReader<Person>();
    reader.setLineMapper(new DefaultLineMapper() {
        {
            setLineTokenizer(new DelimitedLineTokenizer() {
                {
                    setNames(new String[]{"Id","postCode"});
                }
                {
                    setDelimiter("|");
                }
            });
            setFieldSetMapper(new BeanWrapperFieldSetMapper<Person>() {
                {
                    setTargetType(Person.class);
                }
            });
        }
    });
    return reader;
}

Answer 1

Take a look at the PatternMatchingCompositeLineTokenizer . There, you can use a Pattern to identify what records get parsed by what LineTokenizer . In your case, you'd have one Pattern that identifies comma delimited records and map them to the tokenizer that parses via commas. You'd also have a Pattern that identifies records delimited by pipes and maps those to the appropriate LineTokenizer . It would look something like this:

    @Bean
    public LineTokenizer compositeLineTokenizer() throws Exception {
        DelimitedLineTokenizer commaTokenizer = new DelimitedLineTokenizer();

        commaTokenizer.setNames("a", "b", "c");
        commaTokenizer.setDelimiter(",");
        commaTokenizer.afterPropertiesSet();

        DelimitedLineTokenizer pipeTokenizer = new DelimitedLineTokenizer();

        pipeTokenizer.setNames("a", "b", "c");
        pipeTokenizer.setDelimiter("|");

        pipeTokenizer.afterPropertiesSet();

        // I have not tested the patterns here so they may need to be adjusted
        Map<String, LineTokenizer> tokenizers = new HashMap<>(2);
        tokenizers.put("*,*", commaTokenizer);
        tokenizers.put("*|*", pipeTokenizer);

        PatternMatchingCompositeLineTokenizer lineTokenizer = new PatternMatchingCompositeLineTokenizer();

        lineTokenizer.setTokenizers(tokenizers);

        return lineTokenizer;
    }