Convert a Unicode string to an escaped ASCII string

Question

How can I convert this string:

This string contains the Unicode character Pi(π)

into an escaped A

Answer 1

To store actual Unicode codepoints, you have to first decode the String's UTF-16 codeunits to UTF-32 codeunits (which are currently the same as the Unicode codepoints). Use System.Text.Encoding.UTF32.GetBytes() for that, and then write the resulting bytes to the StringBuilder as needed,i.e.

static void Main(string[] args) 
{ 
    String originalString = "This string contains the unicode character Pi(π)"; 
    Byte[] bytes = Encoding.UTF32.GetBytes(originalString);
    StringBuilder asAscii = new StringBuilder();
    for (int idx = 0; idx < bytes.Length; idx += 4)
    { 
        uint codepoint = BitConverter.ToUInt32(bytes, idx);
        if (codepoint <= 127) 
            asAscii.Append(Convert.ToChar(codepoint)); 
        else 
            asAscii.AppendFormat("\\u{0:x4}", codepoint); 
    } 
    Console.WriteLine("Final string: {0}", asAscii); 
    Console.ReadKey(); 
}

Answer 2

A small patch to @Adam Sills's answer which solves FormatException on cases where the input string like "c:\u00ab\otherdirectory\" plus RegexOptions.Compiled makes the Regex compilation much faster:

    private static Regex DECODING_REGEX = new Regex(@"\\u(?<Value>[a-fA-F0-9]{4})", RegexOptions.Compiled);
    private const string PLACEHOLDER = @"#!#";
    public static string DecodeEncodedNonAsciiCharacters(this string value)
    {
        return DECODING_REGEX.Replace(
            value.Replace(@"\\", PLACEHOLDER),
            m => { 
                return ((char)int.Parse(m.Groups["Value"].Value, NumberStyles.HexNumber)).ToString(); })
            .Replace(PLACEHOLDER, @"\\");
    }

Answer 3

Here is my current implementation:

public static class UnicodeStringExtensions
{
    public static string EncodeNonAsciiCharacters(this string value) {
        var bytes = Encoding.Unicode.GetBytes(value);
        var sb = StringBuilderCache.Acquire(value.Length);
        bool encodedsomething = false;
        for (int i = 0; i < bytes.Length; i += 2) {
            var c = BitConverter.ToUInt16(bytes, i);
            if ((c >= 0x20 && c <= 0x7f) || c == 0x0A || c == 0x0D) {
                sb.Append((char) c);
            } else {
                sb.Append($"\\u{c:x4}");
                encodedsomething = true;
            }
        }
        if (!encodedsomething) {
            StringBuilderCache.Release(sb);
            return value;
        }
        return StringBuilderCache.GetStringAndRelease(sb);
    }


    public static string DecodeEncodedNonAsciiCharacters(this string value)
      => Regex.Replace(value,/*language=regexp*/@"(?:\\u[a-fA-F0-9]{4})+", Decode);

    static readonly string[] Splitsequence = new [] { "\\u" };
    private static string Decode(Match m) {
        var bytes = m.Value.Split(Splitsequence, StringSplitOptions.RemoveEmptyEntries)
                .Select(s => ushort.Parse(s, NumberStyles.HexNumber)).SelectMany(BitConverter.GetBytes).ToArray();
        return Encoding.Unicode.GetString(bytes);
    }
}

This passes a test:

public void TestBigUnicode() {
    var s = "\U00020000";
    var encoded = s.EncodeNonAsciiCharacters();
    var decoded = encoded.DecodeEncodedNonAsciiCharacters();
    Assert.Equals(s, decoded);
}

with the encoded value: "\ud840\udc00"

This implementation makes use of a StringBuilderCache (reference source link)