CODEWITHSUNDEEP
javafileio
Robert Smith
Robert Smith

Reputation: 1

Java: Writing/Reading a char to a file yields different results

I'm trying to write a simple character to a file and read it back in. Writing the character to the file appears to work fine (at least as it appears in a hex editor). When I read the character back into memory, its a completely different value altogether. Here's my example code:

public class myclass {

public static void main(String[] args) {
      char myChar = 158; // let myChar = 158

      System.out.println("myChar = "+(int)myChar); // prints 158. Good.   

        try {
            FileOutputStream fileOut = new FileOutputStream("readthis");
                fileOut.write(myChar);
            fileOut.close();
        } catch (IOException e) {
            System.exit(1);
        }


        // If I examine the "readthis" file, there is one byte that has a value of
        // of '9E' or 158. This is what I'd expect.   

        // Lets try to now read it back into memory   


        char readChar = 0;

        try {
            int i = 0;

            FileInputStream fstream = new FileInputStream("readthis");
            DataInputStream in = new DataInputStream(fstream);
            BufferedReader br = new BufferedReader(new InputStreamReader(in));

                readChar = (char)br.read();                     


            in.close();

        } catch (IOException e) {
            System.exit(1);
        }

        // Now, if we look at readChar, it's some value that's not 158!
        // Somehow it got read into as 382!   

        // Printing this value results in 382
        System.out.println("readChar = "+(int)readChar);




  }

}

My question is, how did this happen? I would like readChar to equal its original value that I wrote (158), but I'm not sure what I'm doing wrong. Any help would be appreciated. Thanks.

Upvotes: 0

Views: 414

Answers (3)

Richard LIANG
Richard LIANG

Reputation: 171

If you only want to write/read chars, please try DataOutputStream#writeChar() and DataInputStream#readChar(), but InputStreamRead/OutputStreamWriter is more flexible.

Upvotes: 1

Aaron Digulla
Aaron Digulla

Reputation: 328714

EJP is right. The longer explanation: The character has two properties and you're omitting one: The encoding.

This means that char myChar = 158 assigns myChar the Unicode code point 158 (this isn't a printable character in Unicode).

When you write that to a file as a byte (using fileOut.write(int)), you're converting the Unicode character to the integer 158 - the encoding is lost. The write() method will strip anything but the lower 8 bits from the integer (write(158+256) yields the same result as write(158)).

When you read the data in again, you're using a Reader which reads bytes and converts them into Unicode characters. To do this correctly, you need to specify the encoding with which the data was written. Since you didn't specify anything explicitly, Java uses the platform default encoding (the default for your OS).

So the reader reads 158, and uses the default encoding to turn that into a char.

To fix this, always use Reader/Writer along with InputStreamReader and OutputStreamWriter which allow you to specify which encoding to use. UTF-8 is a good choice since all Java VMs can read them and all Unicode characters can be translated into/from this encoding.

Upvotes: 3

user207421
user207421

Reputation: 311008

You are writing bytes and reading chars. Use a Writer and a Reader, or an OutputStream and an InputStream.

Upvotes: 5

Related Questions