.net - remove - string normalize c#




How do I remove diacritics(accents) from a string in.NET? (12)

I needed something that converts all major unicode characters and the voted answer leaved a few out so I've created a version of CodeIgniter's convert_accented_characters($str) into C# that is easily customisable:

using System;
using System.Text;
using System.Collections.Generic;

public static class Strings
{
    static Dictionary<string, string> foreign_characters = new Dictionary<string, string>
    {
        { "äæǽ", "ae" },
        { "öœ", "oe" },
        { "ü", "ue" },
        { "Ä", "Ae" },
        { "Ü", "Ue" },
        { "Ö", "Oe" },
        { "ÀÁÂÃÄÅǺĀĂĄǍΑΆẢẠẦẪẨẬẰẮẴẲẶА", "A" },
        { "àáâãåǻāăąǎªαάảạầấẫẩậằắẵẳặа", "a" },
        { "Б", "B" },
        { "б", "b" },
        { "ÇĆĈĊČ", "C" },
        { "çćĉċč", "c" },
        { "Д", "D" },
        { "д", "d" },
        { "ÐĎĐΔ", "Dj" },
        { "ðďđδ", "dj" },
        { "ÈÉÊËĒĔĖĘĚΕΈẼẺẸỀẾỄỂỆЕЭ", "E" },
        { "èéêëēĕėęěέεẽẻẹềếễểệеэ", "e" },
        { "Ф", "F" },
        { "ф", "f" },
        { "ĜĞĠĢΓГҐ", "G" },
        { "ĝğġģγгґ", "g" },
        { "ĤĦ", "H" },
        { "ĥħ", "h" },
        { "ÌÍÎÏĨĪĬǏĮİΗΉΊΙΪỈỊИЫ", "I" },
        { "ìíîïĩīĭǐįıηήίιϊỉịиыї", "i" },
        { "Ĵ", "J" },
        { "ĵ", "j" },
        { "ĶΚК", "K" },
        { "ķκк", "k" },
        { "ĹĻĽĿŁΛЛ", "L" },
        { "ĺļľŀłλл", "l" },
        { "М", "M" },
        { "м", "m" },
        { "ÑŃŅŇΝН", "N" },
        { "ñńņňʼnνн", "n" },
        { "ÒÓÔÕŌŎǑŐƠØǾΟΌΩΏỎỌỒỐỖỔỘỜỚỠỞỢО", "O" },
        { "òóôõōŏǒőơøǿºοόωώỏọồốỗổộờớỡởợо", "o" },
        { "П", "P" },
        { "п", "p" },
        { "ŔŖŘΡР", "R" },
        { "ŕŗřρр", "r" },
        { "ŚŜŞȘŠΣС", "S" },
        { "śŝşșšſσςс", "s" },
        { "ȚŢŤŦτТ", "T" },
        { "țţťŧт", "t" },
        { "ÙÚÛŨŪŬŮŰŲƯǓǕǗǙǛŨỦỤỪỨỮỬỰУ", "U" },
        { "ùúûũūŭůűųưǔǖǘǚǜυύϋủụừứữửựу", "u" },
        { "ÝŸŶΥΎΫỲỸỶỴЙ", "Y" },
        { "ýÿŷỳỹỷỵй", "y" },
        { "В", "V" },
        { "в", "v" },
        { "Ŵ", "W" },
        { "ŵ", "w" },
        { "ŹŻŽΖЗ", "Z" },
        { "źżžζз", "z" },
        { "ÆǼ", "AE" },
        { "ß", "ss" },
        { "IJ", "IJ" },
        { "ij", "ij" },
        { "Œ", "OE" },
        { "ƒ", "f" },
        { "ξ", "ks" },
        { "π", "p" },
        { "β", "v" },
        { "μ", "m" },
        { "ψ", "ps" },
        { "Ё", "Yo" },
        { "ё", "yo" },
        { "Є", "Ye" },
        { "є", "ye" },
        { "Ї", "Yi" },
        { "Ж", "Zh" },
        { "ж", "zh" },
        { "Х", "Kh" },
        { "х", "kh" },
        { "Ц", "Ts" },
        { "ц", "ts" },
        { "Ч", "Ch" },
        { "ч", "ch" },
        { "Ш", "Sh" },
        { "ш", "sh" },
        { "Щ", "Shch" },
        { "щ", "shch" },
        { "ЪъЬь", "" },
        { "Ю", "Yu" },
        { "ю", "yu" },
        { "Я", "Ya" },
        { "я", "ya" },
    };

    public static char RemoveDiacritics(this char c){
        foreach(KeyValuePair<string, string> entry in foreign_characters)
        {
            if(entry.Key.IndexOf (c) != -1)
            {
                return entry.Value[0];
            }
        }
        return c;
    }

    public static string RemoveDiacritics(this string s) 
    {
        //StringBuilder sb = new StringBuilder ();
        string text = "";


        foreach (char c in s)
        {
            int len = text.Length;

            foreach(KeyValuePair<string, string> entry in foreign_characters)
            {
                if(entry.Key.IndexOf (c) != -1)
                {
                    text += entry.Value;
                    break;
                }
            }

            if (len == text.Length) {
                text += c;  
            }
        }
        return text;
    }
}

Usage

// for strings
"crème brûlée".RemoveDiacritics (); // creme brulee

// for chars
"Ã"[0].RemoveDiacritics (); // A

I'm trying to convert some strings that are in French Canadian and basically, I'd like to be able to take out the French accent marks in the letters while keeping the letter. (E.g. convert é to e, so crème brûlée would become creme brulee)

What is the best method for achieving this?


I often use an extenstion method based on another version I found here (see Replacing characters in C# (ascii)) A quick explanation:

  • Normalizing to form D splits charactes like è to an e and a nonspacing `
  • From this, the nospacing characters are removed
  • The result is normalized back to form C (I'm not sure if this is neccesary)

Code:

using System.Linq;
using System.Text;
using System.Globalization;

// namespace here
public static class Utility
{
    public static string RemoveDiacritics(this string str)
    {
        if (null == str) return null;
        var chars =
            from c in str.Normalize(NormalizationForm.FormD).ToCharArray()
            let uc = CharUnicodeInfo.GetUnicodeCategory(c)
            where uc != UnicodeCategory.NonSpacingMark
            select c;

        var cleanStr = new string(chars.ToArray()).Normalize(NormalizationForm.FormC);

        return cleanStr;
    }

    // or, alternatively
    public static string RemoveDiacritics2(this string str)
    {
        if (null == str) return null;
        var chars = str
            .Normalize(NormalizationForm.FormD)
            .ToCharArray()
            .Where(c=> CharUnicodeInfo.GetUnicodeCategory(c) != UnicodeCategory.NonSpacingMark)
            .ToArray();

        return new string(chars).Normalize(NormalizationForm.FormC);
    }
}

I've not used this method, but Michael Kaplan describes a method for doing so in his blog post (with a confusing title) that talks about stripping diacritics: Stripping is an interesting job (aka On the meaning of meaningless, aka All Mn characters are non-spacing, but some are more non-spacing than others)

static string RemoveDiacritics(string text) 
{
    var normalizedString = text.Normalize(NormalizationForm.FormD);
    var stringBuilder = new StringBuilder();

    foreach (var c in normalizedString)
    {
        var unicodeCategory = CharUnicodeInfo.GetUnicodeCategory(c);
        if (unicodeCategory != UnicodeCategory.NonSpacingMark)
        {
            stringBuilder.Append(c);
        }
    }

    return stringBuilder.ToString().Normalize(NormalizationForm.FormC);
}

Note that this is a followup to his earlier post: Stripping diacritics....

The approach uses String.Normalize to split the input string into constituent glyphs (basically separating the "base" characters from the diacritics) and then scans the result and retains only the base characters. It's just a little complicated, but really you're looking at a complicated problem.

Of course, if you're limiting yourself to French, you could probably get away with the simple table-based approach in How to remove accents and tilde in a C++ std::string, as recommended by @David Dibben.


In case anyone's interested, here is the java equivalent:

import java.text.Normalizer;

public class MyClass
{
    public static String removeDiacritics(String input)
    {
        String nrml = Normalizer.normalize(input, Normalizer.Form.NFD);
        StringBuilder stripped = new StringBuilder();
        for (int i=0;i<nrml.length();++i)
        {
            if (Character.getType(nrml.charAt(i)) != Character.NON_SPACING_MARK)
            {
                stripped.append(nrml.charAt(i));
            }
        }
        return stripped.toString();
    }
}

It's funny such a question can get so many answers, and yet none fit my requirements :) There are so many languages around, a full language agnostic solution is AFAIK not really possible, as others has mentionned that the FormC or FormD are giving issues.

Since the original question was related to French, the simplest working answer is indeed

    public static string ConvertWesternEuropeanToASCII(this string str)
    {
        return Encoding.ASCII.GetString(Encoding.GetEncoding(1251).GetBytes(str));
    }

1251 should be replaced by the encoding code of the input language.

This however replace only one character by one character. Since I am also working with German as input, I did a manual convert

    public static string LatinizeGermanCharacters(this string str)
    {
        StringBuilder sb = new StringBuilder(str.Length);
        foreach (char c in str)
        {
            switch (c)
            {
                case 'ä':
                    sb.Append("ae");
                    break;
                case 'ö':
                    sb.Append("oe");
                    break;
                case 'ü':
                    sb.Append("ue");
                    break;
                case 'Ä':
                    sb.Append("Ae");
                    break;
                case 'Ö':
                    sb.Append("Oe");
                    break;
                case 'Ü':
                    sb.Append("Ue");
                    break;
                case 'ß':
                    sb.Append("ss");
                    break;
                default:
                    sb.Append(c);
                    break;
            }
        }
        return sb.ToString();
    }

It might not deliver the best performance, but at least it is very easy to read and extend. Regex is a NO GO, much slower than any char/string stuff.

I also have a very simple method to remove space:

    public static string RemoveSpace(this string str)
    {
        return str.Replace(" ", string.Empty);
    }

Eventually, I am using a combination of all 3 above extensions:

    public static string LatinizeAndConvertToASCII(this string str, bool keepSpace = false)
    {
        str = str.LatinizeGermanCharacters().ConvertWesternEuropeanToASCII();            
        return keepSpace ? str : str.RemoveSpace();
    }

And a small unit test to that (not exhaustive) which pass successfully.

    [TestMethod()]
    public void LatinizeAndConvertToASCIITest()
    {
        string europeanStr = "Bonjour ça va? C'est l'été! Ich möchte ä Ä á à â ê é è ë Ë É ï Ï î í ì ó ò ô ö Ö Ü ü ù ú û Û ý Ý ç Ç ñ Ñ";
        string expected = "Bonjourcava?C'estl'ete!IchmoechteaeAeaaaeeeeEEiIiiiooooeOeUeueuuuUyYcCnN";
        string actual = europeanStr.LatinizeAndConvertToASCII();
        Assert.AreEqual(expected, actual);
    }


The CodePage of Greek (ISO) can do it

The information about this codepage is into System.Text.Encoding.GetEncodings(). Learn about in: https://msdn.microsoft.com/pt-br/library/system.text.encodinginfo.getencoding(v=vs.110).aspx

Greek (ISO) has codepage 28597 and name iso-8859-7.

Go to the code... \o/

string text = "Você está numa situação lamentável";

string textEncode = System.Web.HttpUtility.UrlEncode(text, Encoding.GetEncoding("iso-8859-7"));
//result: "Voce+esta+numa+situacao+lamentavel"

string textDecode = System.Web.HttpUtility.UrlDecode(textEncode);
//result: "Voce esta numa situacao lamentavel"

So, write this function...

public string RemoveAcentuation(string text)
{
    return
        System.Web.HttpUtility.UrlDecode(
            System.Web.HttpUtility.UrlEncode(
                text, Encoding.GetEncoding("iso-8859-7")));
}

Note that... Encoding.GetEncoding("iso-8859-7") is equivalent to Encoding.GetEncoding(28597) because first is the name, and second the codepage of Encoding.


This is how i replace diacritic characters to non-diacritic ones in all my .NET program

C#:

//Transforms the culture of a letter to its equivalent representation in the 0-127 ascii table, such as the letter 'é' is substituted by an 'e'
public string RemoveDiacritics(string s)
{
    string normalizedString = null;
    StringBuilder stringBuilder = new StringBuilder();
    normalizedString = s.Normalize(NormalizationForm.FormD);
    int i = 0;
    char c = '\0';

    for (i = 0; i <= normalizedString.Length - 1; i++)
    {
        c = normalizedString[i];
        if (CharUnicodeInfo.GetUnicodeCategory(c) != UnicodeCategory.NonSpacingMark)
        {
            stringBuilder.Append(c);
        }
    }

    return stringBuilder.ToString().ToLower();
}

VB .NET:

'Transforms the culture of a letter to its equivalent representation in the 0-127 ascii table, such as the letter "é" is substituted by an "e"'
Public Function RemoveDiacritics(ByVal s As String) As String
    Dim normalizedString As String
    Dim stringBuilder As New StringBuilder
    normalizedString = s.Normalize(NormalizationForm.FormD)
    Dim i As Integer
    Dim c As Char

    For i = 0 To normalizedString.Length - 1
        c = normalizedString(i)
        If CharUnicodeInfo.GetUnicodeCategory(c) <> UnicodeCategory.NonSpacingMark Then
            stringBuilder.Append(c)
        End If
    Next
    Return stringBuilder.ToString().ToLower()
End Function

Try HelperSharp package.

There is a method RemoveAccents:

 public static string RemoveAccents(this string source)
 {
     //8 bit characters 
     byte[] b = Encoding.GetEncoding(1251).GetBytes(source);

     // 7 bit characters
     string t = Encoding.ASCII.GetString(b);
     Regex re = new Regex("[^a-zA-Z0-9]=-_/");
     string c = re.Replace(t, " ");
     return c;
 }

What this person said:

Encoding.ASCII.GetString(Encoding.GetEncoding(1251).GetBytes(text));

It actually splits the likes of å which is one character (which is character code 00E5, not 0061 plus the modifier 030A which would look the same) into a plus some kind of modifier, and then the ASCII conversion removes the modifier, leaving the only a.



Imports System.Text
Imports System.Globalization

 Public Function DECODE(ByVal x As String) As String
        Dim sb As New StringBuilder
        For Each c As Char In x.Normalize(NormalizationForm.FormD).Where(Function(a) CharUnicodeInfo.GetUnicodeCategory(a) <> UnicodeCategory.NonSpacingMark)  
            sb.Append(c)
        Next
        Return sb.ToString()
    End Function




diacritics