java - 使用opencv抛出异常将字符串转换为pojo类

Question

我需要将我的 http 响应转换为模型，我的响应是纯 csv 格式：

try {
    StringReader stringReader = new StringReader(string);
    CSVReader csvReader = new CSVReaderBuilder(stringReader).build();

    CsvToBean<SampleModel> csvToBean = new CsvToBeanBuilder<SampleModel>(csvReader)
            .withType(SampleModel.class)
            .withIgnoreLeadingWhiteSpace(true)
            .build();
    arrayList =(ArrayList<SampleModel>)csvToBean.parse();
} catch (Exception e) {
    System.out.println(e.getMessage());
}

这让我异常：

捕获 CSV 标头时出错！

字符串如下所示：

Name, Date
Alex, 2012-10-30
Borhan, 2012-11-05

我一直在关注这里的工作：https ://sourceforge.net/p/opencsv/bugs/196/#ee18

标签： javaopencsv

Answer 1

If you have one or more constructors with arguments, create a no-arg constructor explicitly in the bean class. If you do not have one or more constructors with arguments, you do not need to create a no-arg constructor explicitly.

SampleModel.java:

import com.opencsv.bean.CsvBindByName;

public class SampleModel {
    @CsvBindByName(column = "Name")
    private String name;

    @CsvBindByName(column = "Date")
    private String date;

    public SampleModel() {
        // Empty
    }

    public SampleModel(String name, String date) {
        this.name = name;
        this.date = date;
    }

    public void setName(String name) {
        this.name = name;
    }

    public void setDate(String date) {
        this.date = date;
    }

    @Override
    public String toString() {
        return "{" + name + ", " + date + "}";
    }
}

As you can see, I have a two-args constructor and therefore I need to provide a no-arg constructor explicitly. If you remove the no-arg constructor and run Main.java (given below), it will fail at runtime.

However, the following definition of the bean does not require an explicit no-arg constructor. You can verify it by running Main.java again with this definition.

import com.opencsv.bean.CsvBindByName;

public class SampleModel {
    @CsvBindByName(column = "Name")
    private String name;

    @CsvBindByName(column = "Date")
    private String date;

    public void setName(String name) {
        this.name = name;
    }

    public void setDate(String date) {
        this.date = date;
    }

    @Override
    public String toString() {
        return "{" + name + ", " + date + "}";
    }
}

Main.java:

import java.io.StringReader;
import java.util.List;

import com.opencsv.CSVReader;
import com.opencsv.CSVReaderBuilder;
import com.opencsv.bean.CsvToBeanBuilder;

public class Main {
    public static void main(String[] args) {
        String string = "Name, Date\n" + 
                        "Alex, 2012-10-30\n" + 
                        "Borhan, 2012-11-05";

        try {
            StringReader stringReader = new StringReader(string);
            CSVReader csvReader = new CSVReaderBuilder(stringReader).build();
            List<SampleModel> arrayList= new CsvToBeanBuilder<SampleModel>(csvReader)
                                    .withType(SampleModel.class)
                                    .withIgnoreLeadingWhiteSpace(true)
                                    .withSkipLines(1)// Skip the header line
                                    .build()
                                    .parse();
            System.out.println(arrayList);
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
}

Output:

[{Alex,  2012-10-30}, {Borhan,  2012-11-05}]

Original answer:

A workaround is to create an instance of the bean out of each record you read from the CSV and add the instance to the List as shown below:

List<SampleModel> arrayList = new ArrayList<>();
String[] nextRecord;
CSVReader csvReader = new CSVReaderBuilder(stringReader)
                        .withSkipLines(1)// Skip the header line
                        .build();

while ((nextRecord = csvReader.readNext()) != null) {
    arrayList.add(new SampleModel(nextRecord[0], nextRecord[1]));
}