How to find out if string contains non-alpha numeric characters in C#/.NET 2.0?

Question

Allowed characters are (at least) A-Z, a-z, 0-9, ö, Ö, ä, ä, å, Å and german, latvian, estonian (if any) special chars? Is there ready-made method or do i have to make blacklist (non-allowed chars) and regular expressions IsMatch? If no ready-made how to use blacklist?

Answer 1

I don't know how special characters from all those languages are categorised, but you could check if the Char.IsLetterOrDigit method matches what you want to do. It works at least for the digits and letters I tested:

string test = "Aasdf345ÅÄÖåäöéÉóÓüÜïÏôÔ";
if (test.All(Char.IsLetterOrDigit)) { ... }

The Char.IsLetterOrDigit returns true for characters that are categorised in Unicode as UppercaseLetter, LowercaseLetter, TitlecaseLetter, ModifierLetter, OtherLetter, or DecimalDigitNumber.

Answer 2

Investigate char.IsLetterOrDigit(char).

For example:

myString.All(c => char.IsLetterOrDigit(c));

Answer 3

A blacklist for characters is likely pretty large :-)

You can use the regular expression

^[\d\p{L}]+$

to match decimal digits and letters, regardless of script.

This regular expression consists of a character class containing the shorthands \d – which contains every digit (230 in total in the BMP) and \p{L} which contains every Unicode character classified as a "letter" (46817 in the BMP). Said character class is then repeated at least once and embedded between ^ and $ – the string start and end anchors, so it matches the complete string.

For some regex engines, since you're only interested in Latin letters, apparently, you could also use

^[\d\p{Letter}]+$

However, .NET doesn't support this. The first regex mentioned above actually catches everything that's a digit or a letter in any script. So it will dutifully match on Indian or Arabic numerals and Hebrew, Cyrillic and other non-Latin scripts. Depending on what you want this may not be appropriate.

If that poses a problem, then I see no better option than to explicitly list the characters you want to allow. However, I consider it dangerous to assume that text in a certain language is always restricted to that language's script. If I were to write a Czech or Polish name in a German text, then I'd likely need more than just [a-zA-ZäöüÄÖÜß].