有一个UTF-8编码的文本文件,用FileReader读取到一个字符串,然后转换字符集:str=newString(str.getBytes(),"UTF-8");结果大部分中文显示正常,但最后仍有部分汉字显示为问号! 
Java​​​代码​​​  
public static List<String> getLines(String fileName){   
        List<String> lines=newArrayList<String>();   
        try {   
           BufferedReader br = new BufferedReader(new FileReader(fileName));   
            String line= null;   
            while ((line= br.readLine()) != null) {   
               lines.add(newString(line.getBytes("GBK"),"UTF-8"));   
           }   
           br.close();   
        } catch (FileNotFoundException e){   
        }catch (IOException e){}   
        return lines;   
    } 

public staticList<String> getLines(String fileName){
  List<String> lines=new ArrayList<String>();
  try {
   BufferedReader br = new BufferedReader(newFileReader(fileName));
   String line = null;
   while ((line = br.readLine()) != null) {
    lines.add(newString(line.getBytes("GBK"),"UTF-8"));
   }
   br.close();
  } catch (FileNotFoundException e) {
  }catch (IOException e) {}
  return lines;
 }

 

文件读入时是按OS的默认字符集即GBK解码的,我先用默认字符集GBK编码str.getBytes(“GBK”),此时应该还原为文件中的字节序列了,然后再按UTF-8解码,生成的字符串按理说应该就应该是正确的。

为什么结果中还是有部分乱码呢? 
问题出在FileReader读取文件的过程中,FileReader继承了InputStreamReader,但并没有实现父类中带字符集参数的构造函数,所以FileReader只能按系统默认的字符集来解码,然后在UTF-8
-> GBK-> UTF-8的过程中编码出现损失,造成结果不能还原最初的字符。

原因明确了,这个问题解决起来并不困难,用InputStreamReader代替FileReader,InputStreamReaderisr=new
InputStreamReader(new FileInputStream(fileName),"UTF-8");这样读取文件就会直接用UTF-8解码,不用再做编码转换。 
Java代码  
public static List<String> getLines(String fileName){   
        List<String> lines=newArrayList<String>();   
        try {   
           BufferedReader br=new BufferedReader(new InputStreamReader(newFileInputStream(fileName),"UTF-8"));   
            String line= null;   
            while ((line= br.readLine()) != null) {   
               lines.add(line);   
           }   
           br.close();   
        } catch (FileNotFoundException e){   
        }catch (IOException e){}   
        return lines;   
    }