JavaCSVTobean。解析失败，解析CSV时出错

Question

我试图在Eclipse中用OpenCSV解析一个大的CSV文件。 以下是CSV文件的前4条记录。完整文件有219590条记录：-

0,23,1,0,someone@email.com,"Construction/Contractors/Contractors"
0,43,1,0,someone@email.com,"Engineering/Electrical Engineering/Electrical Engineering"
0,395,1,0,someone@email.com,"Sales/Sales Force Management/Sales Management"
0,398,1,0,someone@email.com,"Sales/Sales Strategy/Sales"

下面是Java代码：-

    File csvFile = new File("data/userattrib2_30day.csv");
    ColumnPositionMappingStrategy strategy = new ColumnPositionMappingStrategy();
    strategy.setType(UserAttribRecord.class);
    String[] columns = new String[] {"userId", "attributeId", "rating", "timestamp", "email", "attributeDesc"};
    strategy.setColumnMapping(columns);

    CSVReader reader = new CSVReader(new FileReader(csvFile));
    CsvToBean<UserAttribRecord> csv = new CsvToBean<UserAttribRecord>();
    List<UserAttribRecord> userAttribList = csv.parse(strategy,reader);

它失败于：-

Exception in thread "main" java.lang.RuntimeException: Error parsing CSV!
at com.opencsv.bean.CsvToBean.parse(CsvToBean.java:95)
at com.opencsv.bean.CsvToBean.parse(CsvToBean.java:75)

如何找出错误是什么？似乎没有任何调试信息，所以我不知道哪个字段的哪个记录出现了错误。如何调试这个？ 谢谢

Answer 1

我刚刚尝试了univocity-parsers，我可以毫无问题地解析您的示例输入。试一试，因为它比OpenCSV快两倍（平均快两倍）。这是我的代码：

首先，将@Parsed注释添加到要从CSV加载的字段中（有许多选项可用，请查看文档）

public static class UserAttributeRecord{

    @Parsed
    int userId;

    @Parsed
    int attributeId;

    @Parsed
    int rating;

    @Parsed
    long timestamp;

    @Parsed
    String email;

    @Parsed
    String attributeDesc;
}

以下是解析文件所需的代码：

public static void main(String[] args) throws IOException{

    //creates a processor of java beans.
    BeanListProcessor<UserAttributeRecord> beanProcessor = new BeanListProcessor<UserAttributeRecord>(UserAttributeRecord.class);

    //then a settings object to configure the parser
    CsvParserSettings settings = new CsvParserSettings(); //many options here, check the tutorial

    //configures the parser to use the bean processor.
    settings.setRowProcessor(beanProcessor);

    //configures the input format.
    settings.setHeaders("userId", "attributeId", "rating", "timestamp", "email", "attributeDesc");
    settings.getFormat().setLineSeparator("\n");

    //creates a parser with your settings
    CsvParser parser = new CsvParser(settings);

    //parses everything. All rows are submitted to the row processor defined above
    parser.parseAll(new FileReader(new File("/path/to/file.csv")));

    //here's your list of beans
    List<UserAttributeRecord> beans = beanProcessor.getBeans();
}

披露：我是这个图书馆的作者。它是开源和免费的（Apache V2.0许可证）

Answer 2

如果你能自己构建，那就用opencsv的主干构建吧。我修改了CsvToBean以打印出错误发生时它所在的行号

throw new RuntimeException("Error parsing CSV line: " + lineProcessed + " values: " + Arrays.toString(line), e);

这将在3.6版本中发布，除非出现任何问题，否则应该在感恩节前发布

我很好奇是什么错误导致了这一点。如果你再往下看一下你的异常堆栈跟踪，你应该会看到一个“由引起的：”就是CsvToBean捕获的异常。考虑到你有超过20万行代码，我怀疑你遇到了内存不足的异常——这意味着你要么需要将文件拆分成更小的文件，增加JVM的内存，要么使用3.5版本中引入的IterableCsvToBean，它允许逐行解析

希望有帮助

：）