Thread: FileReader.read() reads garbage

Hi,

I have a problem with the FileReader class. I want to read and copy a file and skip certain bytes. However, when I have a FileReader fr and I have it read(), at a very certain point it reads something else entirely, and subsequently writes that garbage into the new file. It only happens with a certain file though. Here is example code and the file in question:

import java.io.File;
import java.io.FileReader;
import java.io.FileWriter;
 
 
public class Trunc {
 
	/**
	 * @param args
	 * @throws Exception 
	 */
	public static void main(String[] args) throws Exception {
		String in = "in.txt";
		String out = "out.txt";
 
		File inFile = new File(in);
		File outFile = new File(out);
 
		FileReader fr = new FileReader(inFile);
		FileWriter fw = new FileWriter(outFile);
 
		while(fr.ready()) {
                                                int temp = fr.read();
			fw.write(temp);
                                }
 
		fr.close();
		fw.close();
                 }
}

That's really just a fancy way of copying a file byte for byte so I can edit or leave out bytes as they are read.

The problem is this:

The in.txt that I have appended has the value 0x90 at address 0x1E5. However, when I run the above code on in.txt, the out.txt suddenly has the value 0x3F at the same address, and a System.out.print tells me that the read() function has read the nonsense value 0xFFFD at that position!

The same thing happens several times in the file, with different in-values between 0x80 and 0x9D, maybe more, but always replacing the original value with 0x3F in the out file.

It also happens with 1-byte files that contain, for example, the value 0x90 as the only byte. It gets replaced to 0x3f in the out file.

What is the problem here and how can I fix that? Remember that I don't just want to copy the file wholesome, I want to edit/skip certain bytes before writing.

Take a look at the API for FileReader. In particular note the following line:
"FileReader is meant for reading streams of characters. For reading streams of raw bytes, consider using a FileInputStream."

Also consider the character encoding. 0x90 is not in your default character set.
The FileReader API doc says:

constructors of this class assume that the default character encoding

The 0x90 is converted to a ? (0x3F)

To read other character sets, look at setting the default character set.
See the Charset class.

Ok thanks, will try that. Didn't read FileReader properly because I skipped to Reader for the inherited methods.

FileInputStream indeed solved the problem. Thanks again.