Question

if I have a byte [] which holds string in this format:

abcd 546546545 dfdsfdsfd 5415645

and I know the numbers are of type integer. What is the best way to get a raw byte[] out of it, wihtout using String.split() method?

Answer 1

This answer is based on the following assumptions (none of which are clearly warranted from what you have posted):

• You are currently reading bytes directly out of a file
• The file is stored in your VM’s default encoding
• You want to ignore everything that is not a decimal digit
• You want to generate a byte[] where each byte contains the numeric value corresponding to the decimal digits found in the file

With these assumptions, I would solve this problem as follows:

public byte[] getDigitValues(String file) throws IOException {
    FileReader rdr = new FileReader(file);
    ByteArrayOutputStream bos = new ByteArrayOutputStream();
    try {
        rdr = new BufferedReader(rdr);
        for (char c = rdr.read(); c != -1; c = rdr.read()) {
            if (c >= '0' && c <= '9') {
                bos.write(c - '0');
            }
        }
    } finally {
        if (rdr != null) {
            try { rdr.close(); }
            catch (IOException e) {
                throw new IOException("Could not close file", e);
            }
        }
    }
    return bos.toByteArray();
}

In Java 7, I’d use the try-with-resources statement:

public byte[] getDigitValues(String file) throws IOException {
    ByteArrayOutputStream bos = new ByteArrayOutputStream();
    try (Reader rdr = new BufferedReader(new FileReader(file))) {
        for (. . .
    }
    return bos.toByteArray();
}