Should I be using Regex to uppercase uncommon characters?

Question

I've written a function for converting strings to upper case. It currently works by replacing each character using a Regex pattern with a hash:

# Special upcase function that handles several Norwegian symbols.
def norwegian_upcase(string)
    string.upcase.gsub(/[æøåäâãàáëêèéïîìíöôõòóüûùúý]/, {
        'æ' => 'Æ',
        'ø' => 'Ø',
        'å' => 'Å',
        'ä' => 'Ä',
        'â' => 'Â',
        'ã' => 'Ã',
        'à' => 'À',
        'á' => 'Á',
        'ë' => 'Ë',
        'ê' => 'Ê',
        'è' => 'È',
        'é' => 'É',
        'ï' => 'Ï',
        'î' => 'Î',
        'ì' => 'Ì',
        'í' => 'Í',
        'ö' => 'Ö',
        'ô' => 'Ô',
        'õ' => 'Õ',
        'ò' => 'Ò',
        'ó' => 'Ó',
        'ü' => 'Ü',
        'û' => 'Û',
        'ù' => 'Ù',
        'ú' => 'Ú',
        'ý' => 'Ý'
    })
end

I have the feeling that it's horribly inefficient to be using Regex like this, and that I should probably be using another method. Can anyone suggest a better way to replace single characters, or is Regex fit for the task?

Answer 1

One-to-one substitution is precisely what tr is for.

def norwegian_upcase(string)
  string.upcase.tr('æ-ý','Æ-Ý')
end

That being said, the unicode_utils gem provides a method for this:

def norwegian_upcase(string)
  UnicodeUtils.upcase(string, :no)
end

I don't know Norwegian, but you can supply language subtags :nb (Norwegian Bokmal) or :nn (Norwegian Nynorsk) if they would behave differently than general :no (Norwegian).

Answer 2

After a cursory glance at the character codes for these, it looks like the lowercase is 32 (decimal) higher than uppercase. e.g. 'é'.ord - 32 == 'É'.ord

You could try something like this:

string.upcase.gsub(/[æøåäâãàáëêèéïîìíöôõòóüûùúý]/){|s| (s.ord-32).chr(Encoding::UTF_8)}