Why doesn't Đ get flattened to D when Removing Accents/Diacritics

Question

I'm using this method to remove accents from my strings:

static string RemoveAccents(string input)
{
    string normalized = input.Normalize(NormalizationForm.FormKD);
    StringBuilder builder = new StringBuilder();
    foreach (char c in normalized)
    {
        if (char.GetUnicodeCategory(c) !=
        UnicodeCategory.NonSpacingMark)
        {
            builder.Append(c);
        }
    }
    return builder.ToString();
}

but this method leaves đ as đ and doesn't change it to d, even though d is its base char. you can try it with this input string "æøåáâăäĺćçčéęëěíîďđńňóôőöřůúűüýţ"

What's so special in letter đ?

Answer 1

The answer for why it doesn't work is that the statement that "d is its base char" is false. U+0111 (LATIN SMALL LETTER D WITH STROKE) has Unicode category "Letter, Lowercase" and has no decomposition mapping (i.e., it doesn't decompose to "d" followed by a combining mark).

"đ".Normalize(NormalizationForm.FormD) simply returns "đ", which is not stripped out by the loop because it is not a non-spacing mark.

A similar issue will exist for "ø" and other letters for which Unicode provides no decomposition mapping. (And if you're trying to find the "best" ASCII character to represent a Unicode letter, this approach won't work at all for Cyrillic, Greek, Chinese or other non-Latin alphabets; you'll also run into problems if you wanted to transliterate "ß" into "ss", for example. Using a library like UnidecodeSharp may help.)

Answer 2

I have to admit that I'm not sure why this works but it sure seems to

var str = "æøåáâăäĺćçčéęëěíîďđńňóôőöřůúűüýţ";
var noApostrophes = Encoding.ASCII.GetString(Encoding.GetEncoding("Cyrillic").GetBytes(str)); 

=> "aoaaaaalccceeeeiiddnnooooruuuuyt"

Answer 3

"D with stroke" (Wikipedia) is used in several languages, and appears to be considered a distinct letter in all of them -- and that is why it remains unchanged.