Regex with IgnoreCase option may match more characters than expected

This post is part of the series 'Strings in .NET'. Be sure to check out the rest of the blog posts of the series!

• String comparisons are harder than it seems
• .NET Regex: \d is different from [0-9]
• How to correctly count the number of characters of a string
• Correctly converting a character to lower/upper case
• How not to read a string from a UTF-8 stream
• Regex with IgnoreCase option may match more characters than expected (this post)
• How to remove diacritics from a string in .NET

In a previous post, I explained why \d is different from [0-9]. In this post, I'll explain why the regex [a-zA-Z] differs from [a-z] when using the IgnoreCase option.

var regex1 = new Regex("^[a-zA-Z]+$");
var regex2 = new Regex("^[a-z]+$", RegexOptions.IgnoreCase | RegexOptions.CultureInvariant);

Console.WriteLine(regex1.IsMatch("Test")); // true
Console.WriteLine(regex2.IsMatch("Test")); // true

However, you get different results if you use the Kelvin sign:

// Kelvin Sign (U+212A)
Console.WriteLine(regex1.IsMatch("K")); // false
Console.WriteLine(regex2.IsMatch("K")); // true

When the RegexOptions.IgnoreCase option is specified, comparisons between the input and the pattern are case-insensitive. To support this, the Regex engine uses case mappings. A match occurs if two characters are considered equivalent according to the casing rules (e.g., one is the lowercase version of the other).

In this case, char.ToLowerInvariant('K') (Kelvin Sign) returns 'k' (Latin Small Letter K). Therefore, when using the IgnoreCase option, [a-z] matches the Kelvin sign because 'k' is in the range a-z. However, the Kelvin Sign itself is not part of the [a-zA-Z] set, so the regex [a-zA-Z] does not match it.

To conclude, the following regular expressions are equivalent:

new Regex("^[a-z]+$", RegexOptions.IgnoreCase | RegexOptions.CultureInvariant);
new Regex("^[A-Za-z\u212A]+$");

If you are curious about how .NET knows which case mappings to use, you can read the code of the GenerateRegexCasingTable tool on GitHub.

#Additional resources

• Regex: \d is not [0-9]
• RegexOptions.IgnoreCase documentation

Do you have a question or a suggestion about this post? Contact me!