Convert String to UTF Byte Array

Description

Demo Code


//package com.java2s;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.io.OutputStream;

public class Main {
    public static void main(String[] argv) {
        String string = "java2s.com";
        System.out.println(java.util.Arrays
                .toString(toUTF8ByteArray(string)));
    }//w w w . j a va 2 s . c  om

    public static byte[] toUTF8ByteArray(String string) {
        return toUTF8ByteArray(string.toCharArray());
    }

    public static byte[] toUTF8ByteArray(char[] string) {
        ByteArrayOutputStream bOut = new ByteArrayOutputStream();

        try {
            toUTF8ByteArray(string, bOut);
        } catch (IOException e) {
            throw new IllegalStateException(
                    "cannot encode string to byte array!");
        }

        return bOut.toByteArray();
    }

    public static void toUTF8ByteArray(char[] string, OutputStream sOut)
            throws IOException {
        char[] c = string;
        int i = 0;

        while (i < c.length) {
            char ch = c[i];

            if (ch < 0x0080) {
                sOut.write(ch);
            } else if (ch < 0x0800) {
                sOut.write(0xc0 | (ch >> 6));
                sOut.write(0x80 | (ch & 0x3f));
            }
            // surrogate pair
            else if (ch >= 0xD800 && ch <= 0xDFFF) {
                // in error - can only happen, if the Java String class has a
                // bug.
                if (i + 1 >= c.length) {
                    throw new IllegalStateException(
                            "invalid UTF-16 codepoint");
                }
                char W1 = ch;
                ch = c[++i];
                char W2 = ch;
                // in error - can only happen, if the Java String class has a
                // bug.
                if (W1 > 0xDBFF) {
                    throw new IllegalStateException(
                            "invalid UTF-16 codepoint");
                }
                int codePoint = (((W1 & 0x03FF) << 10) | (W2 & 0x03FF)) + 0x10000;
                sOut.write(0xf0 | (codePoint >> 18));
                sOut.write(0x80 | ((codePoint >> 12) & 0x3F));
                sOut.write(0x80 | ((codePoint >> 6) & 0x3F));
                sOut.write(0x80 | (codePoint & 0x3F));
            } else {
                sOut.write(0xe0 | (ch >> 12));
                sOut.write(0x80 | ((ch >> 6) & 0x3F));
                sOut.write(0x80 | (ch & 0x3F));
            }

            i++;
        }
    }

    public static byte[] toByteArray(char[] chars) {
        byte[] bytes = new byte[chars.length];

        for (int i = 0; i != bytes.length; i++) {
            bytes[i] = (byte) chars[i];
        }

        return bytes;
    }

    public static byte[] toByteArray(String string) {
        byte[] bytes = new byte[string.length()];

        for (int i = 0; i != bytes.length; i++) {
            char ch = string.charAt(i);

            bytes[i] = (byte) ch;
        }

        return bytes;
    }
}
Convert String to UTF Byte Array - Java java.lang

Description

Demo Code

Related Tutorials
