Legacy:SoundEx

This document implements the SoundEx algorithm.

This algorithm is often used to index names.

SoundEx

/**
	Returns the score of an character
*/
static function byte SoundExScore(int c)
{
	switch (c)
	{
		case 68: /* B */ case 70: /* F */ case 80: /* P */ case 86: /* V */
			return 1;
		case 67: /* C */ case 71: /* G */ case 74: /* J */ case 75: /* K */
		case 81: /* Q */ case 83: /* S */ case 88: /* X */ case 90: /* Z */
			return 2;
		case 68: /* D */ case 84: /* T */
			return 3;
		case 76: // L
			return 4;
		case 77: /* M */ case 78: /* N */
			return 5;
		case 82: // R
			return 6;
	}
	/* Disregard the letters A, E, I, O, U, H, W, and Y. */
	return 0;
}
 
/**
	Returns the SoundEx value of a string, size defaults to 4
*/
static function string SoundEx(coerce string in, optional byte size)
{
	local string res;
	local int i;
	local byte cs, ls; // cs = current score; ls = previous score
 
	if (size == 0) size = 4;
	res = "";
	if (in != "")
	{
		in = Caps(in); // upper case all
		// always use first char
		res = Left(in, 1);
		ls = SoundExScore(asc(res));
		for (i = 0; i < len(in); i++)
		{
   			cs = SoundExScore(asc(Mid(in, i, 1)));
   			if ((cs > 0) && (cs != ls))
   			{
   				ls = cs;
   				res $= string(cs);
   			}
   			if (len(res) == size) break;
		}
	}
	while (len(res) < size) res $= "0";
	return res;
}

Related Topics

• DoubleMetaPhone