Spring批处理FlatfileitemReader读取格式错误的行

Question

在我的项目中，我正在使用Spring批处理，并使用FlatFileItemReader / FieldSetMapper读取文件。 某些输入文件有问题。行被剪切/格式错误，记录很少。
假设输入文件有4列。 几列格式不正确。 谁能帮我解决这个问题？（如果需要，我可以解释更多）
FILE.CSV

"id","name","age","salary"
"1","user1","28","1000"
"2","user2","27","2000"
"3","user3","26
    ","3000"
"4","user4","25","
    4000"
"5","
        user5","24","5000"
"6","user6","23","6000"
"7","user7","22","7000"
"8","user8","21","8000"

Answer 1

在使用FlatFileItemReader读取格式错误的行时，我遇到了类似的问题。 在这种情况下，可以将DefaultRecordSeparatorPolicy用作FlatFileItemReader中的RecordSeparatorPolicy。 它的作用是在读取一行后检查endOfRecord。 如果读取的行中有任何未注释的引号，它将读取另一行以规范化输入。 您还可以覆盖行为。

flatFileItemReader.setRecordSeparatorPolicy(new DefaultRecordSeparatorPolicy());

有关更多信息，请参考DefaultRecordSeparatorPolicy API文档

@Bean
public FlatFileItemReader<YourClassName> itemReader(@Value("${input}") Resource resource) {
    FlatFileItemReader<YourClassName> flatFileItemReader = new FlatFileItemReader<>();
    flatFileItemReader.setResource(resource);
    flatFileItemReader.setName("CSV-Reader");
    flatFileItemReader.setLinesToSkip(1);
    // override default comment '#' from file parsing
    flatFileItemReader.setComments(new String[] {});
    // checks for multi-line csv inputs for very lage row
    flatFileItemReader.setRecordSeparatorPolicy(new DefaultRecordSeparatorPolicy());
    flatFileItemReader.setLineMapper(lineMapper());
    return flatFileItemReader;
}

@Bean
public LineMapper<YourClassName> lineMapper() {
    DelimitedLineTokenizer lineTokenizer = new DelimitedLineTokenizer();
    lineTokenizer.setDelimiter(DelimitedLineTokenizer.DELIMITER_COMMA);
    lineTokenizer.setQuoteCharacter(DelimitedLineTokenizer.DEFAULT_QUOTE_CHARACTER);
    lineTokenizer.setStrict(false);
    lineTokenizer.setNames(COLUMN_NAMES);

    BeanWrapperFieldSetMapper<YourClassName> fieldSetMapper = new BeanWrapperFieldSetMapper<>();
    fieldSetMapper.setTargetType(YourClassName.class);

    DefaultLineMapper<YourClassName> defaultLineMapper = new DefaultLineMapper<>();
    defaultLineMapper.setLineTokenizer(lineTokenizer);
    defaultLineMapper.setFieldSetMapper(fieldSetMapper);
    return defaultLineMapper;
}