Accidentally splitting unicode chars when truncating strings

user_492067 0
Viewed 667

I'm saving some strings from a third party into my database (postgres). Sometimes these strings are too long and need to be truncated to fit into the column in my table.

On some random occasions I accidentally truncate the string right where there is a Unicode character, which gives me a "broken" string that I cannot save into the database. I get the following error: Unable to translate Unicode character \uD83D at index XXX to specified code page.

I've created a minimal example to show you what I mean. Here I have a string that contains a Unicode character ("Small blue diamond" U+1F539). Depending on where I truncate, it gives me a valid string or not.

var myString = @"This is a string before an emoji: This is after the emoji.";

var brokenString = myString.Substring(0, 34);
// Gives: "This is a string before an emoji:☐"

var test3 = myString.Substring(0, 35);
// Gives: "This is a string before an emoji:"

Is there a way for me to truncate the string without accidentally breaking any Unicode chars?

unicode postgresql string c#
4 Answers

A Unicode character may be represented with several chars, that is the problem with string.Substring you are having.

You may convert your string to a StringInfo object and then use SubstringByTextElements() method to get the substring based on the Unicode character count, not a char count.

See a C# demo:

Console.WriteLine("".Length); // => 2
Console.WriteLine(new StringInfo("").LengthInTextElements); // => 1

var myString = @"This is a string before an emoji:This is after the emoji.";
var teMyString = new StringInfo(myString);
Console.WriteLine(teMyString.SubstringByTextElements(0, 33));
// => "This is a string before an emoji:"
Console.WriteLine(teMyString.SubstringByTextElements(0, 34));
// => This is a string before an emoji:
Console.WriteLine(teMyString.SubstringByTextElements(0, 35));
// => This is a string before an emoji:T
user_38329700

I ended up using a modification of xanatos answer here. The difference is that this version will strip the last grapheme, if adding it would give a string longer than length.

    public static string UnicodeSafeSubstring(this string str, int startIndex, int length)
    {
        if (str == null)
        {
            throw new ArgumentNullException(nameof(str));
        }

        if (startIndex < 0 || startIndex > str.Length)
        {
            throw new ArgumentOutOfRangeException(nameof(startIndex));
        }

        if (length < 0)
        {
            throw new ArgumentOutOfRangeException(nameof(length));
        }

        if (startIndex + length > str.Length)
        {
            throw new ArgumentOutOfRangeException(nameof(length));
        }

        if (length == 0)
        {
            return string.Empty;
        }

        var stringBuilder = new StringBuilder(length);

        var enumerator = StringInfo.GetTextElementEnumerator(str, startIndex);

        while (enumerator.MoveNext())
        {
            var grapheme = enumerator.GetTextElement();
            startIndex += grapheme.Length;

            if (startIndex > str.Length)
            {
                break;
            }

            // Skip initial Low Surrogates/Combining Marks
            if (stringBuilder.Length == 0)
            {
                if (char.IsLowSurrogate(grapheme[0]))
                {
                    continue;
                }

                var cat = char.GetUnicodeCategory(grapheme, 0);

                if (cat == UnicodeCategory.NonSpacingMark || cat == UnicodeCategory.SpacingCombiningMark || cat == UnicodeCategory.EnclosingMark)
                {
                    continue;
                }
            }

            // Do not append the grapheme if the resulting string would be longer than the required length
            if (stringBuilder.Length + grapheme.Length <= length)
            {
                stringBuilder.Append(grapheme);
            }

            if (stringBuilder.Length >= length)
            {
                break;
            }
        }

        return stringBuilder.ToString();
    }
}
user_4920670

Here is an example for truncate (startIndex = 0):

string truncatedStr = (str.Length > maxLength)
    ? str.Substring(0, maxLength - (char.IsLowSurrogate(str[maxLength]) ? 1 : 0))
    : str;
user_22040010

Better truncate by the number of bytes not string length

   public static string TruncateByBytes(this string text, int maxBytes)
    {
        if (string.IsNullOrEmpty(text) || Encoding.UTF8.GetByteCount(text) <= maxBytes)
        {
            return text;
        }
        var enumerator = StringInfo.GetTextElementEnumerator(text);
        var newStr = string.Empty;
        do
        {
            enumerator.MoveNext();
            if (Encoding.UTF8.GetByteCount(newStr + enumerator.Current) <= maxBytes)
            {
                newStr += enumerator.Current;
            }
            else
            {
                break;
            }
        } while (true);
        return newStr;
    }
user_6315270
Related