Copy bytes of String to chosen array in chosen charset

Question

In Java 1.8.0, I would like to have the bytes of a String copied to a given array without a new array being returned. I have found three methods which approximate what I would like:

1. byte[] String.getBytes(String charsetName) gives the result I would like, but it returns a new array.

2. void String.getBytes(int srcBegin, int srcEnd, byte dst[], int dstBegin) copies to a given array, but it has no option for choosing the charset, dropping the upper 8 bits of each char in char[] String.value. This method also happens to be deprecated.

3. byte[] StringCoding.encode(Charset cs, char[] ca, int off, int len), which is called by the above String.getBytes, also does the job, but it returns a new array. This method also happens not to be publicly usable.

My use case is iterating over a collection of strings, reading their bytes into an existing structure which does not care about having a unique byte array for every string. One-time allocation of objects at the beginning is okay, but I do not want to allocate objects while iterating over the strings.

Could anyone suggest how I could have the functionality of the second method without the allocation of a new array or any other object?

Answer 1

You can wrap your array in a ByteBuffer, and use a CharsetEncoder directly:

byte[] byteArray = ...;
String charsetName = ...;

ByteBuffer buffer =
    ByteBuffer.wrap(byteArray, dstBegin, dstEnd - dstBegin);
CharsetEncoder encoder = Charset.forName(charsetName).newEncoder();
CoderResult result =
    encoder.encode(CharBuffer.wrap(string), buffer, true);
if (!result.isUnderflow()) {
    result.throwException();
}

Note that ByteBuffer.wrap is lightweight and does not create a new array; it just provides a ByteBuffer instance that delegates its storage to the byte array given to the method.

If you plan to use the entire byte array, you can just use the easier-to-read ByteBuffer.wrap(byteArray).