Java ByteBuffer to String

Question

Is this a correct approach to convert ByteBuffer to String in this way,

String k = "abcd";
ByteBuffer b = ByteBuffer.wrap(k.getBytes());
String v = new String(b.array());

if(k.equals(v))
    System.out.println("it worked");
else
    System.out.println("did not work");

The reason I ask is that is this looks too simple, whereas other approaches like Java: Converting String to and from ByteBuffer and associated problems looks more complex.

Solution 1

There is simpler approach to decode a ByteBuffer into a String without any problems, mentioned by Andy Thomas.

String s = StandardCharsets.UTF_8.decode(byteBuffer).toString();

Solution 2

EDIT (2018): The edited sibling answer by @xinyongCheng is a simpler approach, and should be the accepted answer.

Your approach would be reasonable if you knew the bytes are in the platform's default charset. In your example, this is true because k.getBytes() returns the bytes in the platform's default charset.

More frequently, you'll want to specify the encoding. However, there's a simpler way to do that than the question you linked. The String API provides methods that converts between a String and a byte[] array in a particular encoding. These methods suggest using CharsetEncoder/CharsetDecoder "when more control over the decoding [encoding] process is required."

To get the bytes from a String in a particular encoding, you can use a sibling getBytes() method:

byte[] bytes = k.getBytes( StandardCharsets.UTF_8 );

To put bytes with a particular encoding into a String, you can use a different String constructor:

String v = new String( bytes, StandardCharsets.UTF_8 );

Note that ByteBuffer.array() is an optional operation. If you've constructed your ByteBuffer with an array, you can use that array directly. Otherwise, if you want to be safe, use ByteBuffer.get(byte[] dst, int offset, int length) to get bytes from the buffer into a byte array.

Solution 3

Try this:

new String(bytebuffer.array(), "ASCII");

NB. you can't correctly convert a byte array to a String without knowing its encoding.

I hope this helps

Solution 4

Just wanted to point out, it's not safe to assume ByteBuffer.array() will always work.

byte[] bytes;
if(buffer.hasArray()) {
    bytes = buffer.array();
} else {
    bytes = new byte[buffer.remaining()];
    buffer.get(bytes);
}
String v = new String(bytes, charset);

Usually buffer.hasArray() will always be true or false depending on your use case. In practice, unless you really want it to work under any circumstances, it's safe to optimize away the branch you don't need. But the rest of the answers may not work with a ByteBuffer that's been created through ByteBuffer.allocateDirect().

Solution 5

The answers referring to simply calling array() are not quite correct: when the buffer has been partially consumed, or is referring to a part of an array (you can ByteBuffer.wrap an array at a given offset, not necessarily from the beginning), we have to account for that in our calculations. This is the general solution that works for buffers in all cases (does not cover encoding):

if (myByteBuffer.hasArray()) {
    return new String(myByteBuffer.array(),
        myByteBuffer.arrayOffset() + myByteBuffer.position(),
        myByteBuffer.remaining());
} else {
    final byte[] b = new byte[myByteBuffer.remaining()];
    myByteBuffer.duplicate().get(b);
    return new String(b);
}

For the concerns related to encoding, see Andy Thomas' answer.

View more solutions

Comments

[tckmn]

Well, did you try it?

[vikky.rk]

Yes I did and it works. But I have seen other implementations which are more complex, like stackoverflow.com/questions/1252468/…

[Gus]

@Doorknob et. al. He's missing encoding and his example (when syntax is corrected) will work, but his method is still not right.

[Gus]

UTF-8 is probably a better default guess than ASCII?

[Andy Thomas]

Neither should be specified, given the OP's use of k.getBytes(), which uses the platform's default charset.

[William Kinaan]

and in the ByteBuffer.get function, the input is again an array of bytes, how can I get it? it doesn't make any sense to say again k.getbytes, does it?

[Andy Thomas]

@WilliamKinaan - You have the byte[] you fed to ByteBuffer.get(byte[] dst, int offset, int length). You you can build a String out of it with the String() constructor `String(byte[] bytes, int offset, int length, Charset charset). You can use the same offset and length values for both calls.

[Dzmitry Lazerka]

Not all buffers are backed by an array, so .array() may throw an exception.

[Madura Pradeep]

There's no k.getBytes() method in java.nio.ByteBuffer (may be not in the version am using). So I used k.array() method which will return byte[].

[Andy Thomas]

@MaduraPradeep - In the example code in the question and this answer, k is a String, not a ByteBuffer.

[Lev Kuznetsov]

If the buffer is created via ByteBuffer.wrap(bytes, offset, size) factory .array() will return the entire bytes array. Better use the form xinyong Cheng suggested

[Fuwjax]

The .decode() on Charset is a better solution, agreed. I do feel the context of my answer is useful information, but much less so now.

[asmaier]

Be aware that UTF-8 might not be the optimal charset for converting bytes to strings and vice-versa. For a 1-to-1 mapping of bytes to chars better use ISO-8859-1, see stackoverflow.com/questions/9098022/…

[asmaier]

Be aware that UTF-8 might not be the optimal charset for converting bytes to strings and vice-versa. For a 1-to-1 mapping of bytes to chars better use ISO-8859-1, see stackoverflow.com/questions/9098022/….

[ScalaWilliam]

Not all bytebuffers support the .array() method.

[Mahesh Pujari]

new String(buffer.array(), 0, length, "UTF-8") good approach would be to specify lengths, start and end

[oliver]

Careful! If you use array(), you must also use arrayOffset() to start at the correct position in the array! This is a subtle pitfall, because usually arrayOffset() is 0; but in those rare cases where it isn't you will get hard-to-find bugs if you don't take it into account.

[oliver]

Careful! If you use array(), you must also use arrayOffset() to start at the correct position in the array! This is a subtle pitfall, because usually arrayOffset() is 0; but in those rare cases where it isn't you will get hard-to-find bugs if you don't take it into account.

[David Ehrmann]

Also, of you don't really need a string, the CharBuffer decode() returns is a CharSequence (like String), so you can avoid an extra copy and use it directly.

[Björn Lindqvist]

Your code is not decoding from latin1 to utf8. While your code is correct, calling the CharBuffer utf8Buffer is somewhat misleading because it has no encoding.

[Dustin]

There is yet another problem with this solution that bit me. The ByteBuffer has an arrayOffset(). Normally, this only happens when you slice() a Buffer, but Android was allocating() buffers with an offset! Extremely hard issue to find.

[Binita Bharati]

This was the only method that worked for me on a java.nio.ByteBuffer instance. String fileContent = new String(bb.array(), StandardCharsets.UTF_8);

[Tom Anderson]

@DavidEhrmann CharBuffer does not have a decode method - what are you referring to here?

[David Ehrmann]

@TomAnderson Charset has a decode() method.

[Tom Anderson]

@DavidEhrmann Ah! I completely misparsed your comment, sorry!