混合情况下base16（十六进制）解码的轻量级方法

Question

Guava has interface that allows for base16/hexadecimal encoding and decoding. Is there a lightweight method or trick for reading in hexadecimals that may contain mixed case?

With lightweight I mean:

• no copying of the input string (so no toUpperCase() allowed)
• few lines of code
• low amount of additional instructions (performant)

The solution should not require an additional library on top of Java SE, and should be able to return relatively large byte arrays.

Code sample:

byte[] decoded = BaseEncoding.base16().decode("abCD");

will result in:

com.google.common.io.BaseEncoding$DecodingException: Unrecognized character: a

using Guava version 15.0 (the latest release at the time of this question).

Answer 1

Integer.parseInt(string, 16); does exactly what you describe. If that's too easy for you, you may look into its implementation to learn how to do such a thing. But it's more than a few lines as it will handle all corner cases you might forget when doing it yourself. Eg the allowed maximum absolute value differs for positive and negative values. The only difference between an optimized hexadecimal and an arbitrary radix parsing is that you can use result<<=4; instead of result *= radix; for the hexadecimal case.

---

Given your updated question the recommendation to look into that method still applies. Decoding a sequence of bytes instead of an int will be even simpler:

import static java.lang.Character.digit;
…

static byte[] decode(CharSequence cs)
{
  final int numCh = cs.length();
  if((numCh&1)!=0)
    throw new IllegalArgumentException("cs must have an even length");
  byte[] array=new byte[numCh>>1];
  for(int p=0; p<numCh; p+=2)
  {
    int hi=digit(cs.charAt(p), 16), lo=digit(cs.charAt(p+1), 16);
    if((hi|lo)<0)
      throw new IllegalArgumentException(cs+" contains non-hex characters");
    array[p>>1]=(byte)(hi<<4|lo);
  }
  return array;
}

Answer 2

I guess Apache commons-codec Hex class contains the methods you want. The actual code for converting chars to digits is:

protected static int toDigit(final char ch, final int index) throws DecoderException {
    final int digit = Character.digit(ch, 16);
    if (digit == -1) {
        throw new DecoderException("Illegal hexadecimal character " + ch + " at index " + index);
    }
    return digit;
}

And the property Character.digit('F', 16) == Character.digit('f', 16) holds.