Removing non alpha characters

Question

What is the best way in order to remove all non-alpha characters in C#? I have looked up Regex but it doesn't seem to recognise Regex when I do:

string cleanString = "";
    string dirtyString = "I don't_8 really know what ! 6 non alpha- is?";
        cleanString = Regex.Replace(dirtyString, "[^A-Za-z0-9]", "");

Regex comes with a red wiggly line underneath. Is there a way I can remove simply non alpha letters and if so can some provide me with a sample? I'm not sure if loops and arrays are the way to go and also how can I get all non alpha characters? I'm assuming I have to do something like if doesn't equal A-Z or 0-9, then remove with ""?

Answer 1

You can do it using LINQ like so:

var cleanString = new string(dirtyString.Where(Char.IsLetter).ToArray());

You can check other Char checks on MSDN.

Answer 2

Regex comes with a red wiggly line underneath.

Then either:

1. The compilation prediction isn't working correctly (it does sometimes get things wrong).
2. You don't have a using System.Text.RegularExpressions in the code, so it can't work out you mean System.Text.RegularExpressions.Regex when you say Regex.

To return to your original question:

What is the best way in order to remove all non-alpha characters in C#?

The approach you take is good for small strings, though [^A-Za-z0-9] will remove non-alphanumerics and [^A-Za-z] non-alphabetical characters. This is assuming you are already restricted to (or want to add a restriction to) US-ASCII characters. To include letters like á, œ, ß or δ because you're dealing with real words rather than computer-code I'd use @"\P{L}" or @"[^\p{L}\p{N}]" to allow all letters and numbers.

If you are dealing with very large piece of text (many kilobytes) then you are better off reading it through a filtering stream that strips the characters you don't want as you go.