Question

I have a hexidecimal string that I need to convert to a byte array. The best way (ie efficient and least code) is:

string hexstr = '683A2134'; byte[] bytes = new byte[hexstr.Length/2]; for(int x = 0; x < bytes.Length; x++) {     bytes[x] = Convert.ToByte(hexstr.Substring(x * 2, 2), 16); } 

In the case where I have a 32bit value I can do the following:

string hexstr = '683A2134'; byte[] bytes = BitConverter.GetBytes(Convert.ToInt32(hexstr, 16));  

However what about in the general case? Is there a better built in function, or a clearer (doesn’t have to be faster, but still performant) way of doing this?

I would prefer a built in function as there seems to be one for everything (well common things) except this particular conversion.

Answer 1

You get the best performance if you calculate the values from the character codes instead of creating substrings and parsing them.

Code in C#, that handles both upper and lower case hex (but no validation):

static byte[] ParseHexString(string hex) {     byte[] bytes = new byte[hex.Length / 2];     int shift = 4;     int offset = 0;     foreach (char c in hex) {         int b = (c - '0') % 32;         if (b > 9) b -= 7;         bytes[offset] |= (byte)(b << shift);         shift ^= 4;         if (shift != 0) offset++;     }     return bytes; } 

Usage:

byte[] bytes = ParseHexString('1fAB44AbcDEf00'); 

As the code uses a few tricks, here a commented version:

static byte[] ParseHexString(string hex) {     // array to put the result in     byte[] bytes = new byte[hex.Length / 2];     // variable to determine shift of high/low nibble     int shift = 4;     // offset of the current byte in the array     int offset = 0;     // loop the characters in the string     foreach (char c in hex) {         // get character code in range 0-9, 17-22         // the % 32 handles lower case characters         int b = (c - '0') % 32;         // correction for a-f         if (b > 9) b -= 7;         // store nibble (4 bits) in byte array         bytes[offset] |= (byte)(b << shift);         // toggle the shift variable between 0 and 4         shift ^= 4;         // move to next byte         if (shift != 0) offset++;     }     return bytes; }