return only Digits 0-9 from a String

I need a regular expression that I can use in VBScript and .NET that will return only the numbers that are found in a string.

For Example any of the following "strings" should return only 1231231234

• 123 123 1234
• (123) 123-1234
• 123-123-1234
• (123)123-1234
• 123.123.1234
• 123 123 1234
• 1 2 3 1 2 3 1 2 3 4

This will be used in an email parser to find telephone numbers that customers may provide in the email and do a database search.

I may have missed a similar regex but I did search on regexlib.com.

[EDIT] - Added code generated by RegexBuddy after setting up musicfreak's answer

VBScript Code

Dim myRegExp, ResultString
Set myRegExp = New RegExp
myRegExp.Global = True
myRegExp.Pattern = "[^\d]"
ResultString = myRegExp.Replace(SubjectString, "")

VB.NET

Dim ResultString As String
Try
      Dim RegexObj As New Regex("[^\d]")
      ResultString = RegexObj.Replace(SubjectString, "")
Catch ex As ArgumentException
      'Syntax error in the regular expression
End Try

C#

string resultString = null;
try {
    Regex regexObj = new Regex(@"[^\d]");
    resultString = regexObj.Replace(subjectString, "");
} catch (ArgumentException ex) {
    // Syntax error in the regular expression
}

From stackoverflow

• Have you gone through the phone nr category on regexlib. Seems like quite a few do what you need.

• By the looks of things, your trying to catch any 10 digit phone number....

  Why not do a string replace first of all on the text to remove any of the following characters.

  <SPACE> , . ( ) - [ ]
  

  Then afterwards, you can just do a regex search for a 10 digit number.

  \d{10}
  

  Brian Boatright : that is what's in place but I wanted to make it match a wider range of input string.

• I don't know if VBScript has some kind of a "regular expression replace" function, but if it does, then you could do something like this pseudocode:

  reg_replace(/\D+/g, '', your_string)
  

  I don't know VBScript so I can't give you the exact code but this would remove anything that is not a number.

  EDIT: Make sure to have the global flag (the "g" at the end of the regexp), otherwise it will only match the first non-number in your string.

  Brian Boatright : Thanks! That's exactly what I was looking to do. I knew it had to be somewhat simple. I'm using RegExBuddy and will try to test it and then post the VBScript code. I believe VBScript will do a replace.

  Matthew Flaschen : If you want to do it with .NET classes, it's basically re = Regex("\D"); re.Replace("123 123 1234", ""). Remember to cache your Regex objects (don't compile them every time the method is called).

• Just installing C# Express so I can test this code, but in .NET, couldn't you simply extract just the digits from the string? Something like this:

  string justNumbers = new String(text.Where(Char.IsDigit).ToArray());
  

  Brian Boatright : that's very cool.

  Matt Hamilton : ps. I know I've answered a VB question with C#, but since it's .NET I figured it's worth putting the idea out there. RegEx seems like overkill for something this simple.

  Brian Boatright : I actually needed VBScript to use in a Classic ASP page but I appreciate your answer.

  Matthew Flaschen : I was about to post a comment along the lines of, "/Clearly/, regex would be faster for this", but I ran a (unscientific) benchmark in Mono, and Linq won (about half the duration the regex took). :) So my hat is off to you.

  Mohamed : that is elegant piece of code.

• Note: you've only solved half the problem here.

  For US phone numbers entered "in the wild", you may have:

  • Phone numbers with or without the "1" prefix
  • Phone numbers with or without the area code
  • Phone numbers with extension numbers (if you blindly remove all non-digits, you'll miss the "x" or "Ext." or whatever also on the line).
  • Possibly, numbers encoded with mnemonic letters (800-BUY-THIS or whatever)

  You'll need to add some smarts to your code to conform the resulting list of digits to a single standard that you actually search against in your database.

  Some simple things you could do to fix this:

  • Before the RegEx removal of non-digits, see if there's an "x" in the string. If there is, chop everything off after it (will handle most versions of writing an extension number).

  • For any number with 10+ digits beginning with a "1", chop off the 1. It's not part of the area code, US area codes start in the 2xx range.

  • For any number still exceeding 10 digits, assume the remainder is an extension of some sort, and chop it off.

  • Do your database search using an "ends-with" pattern search (SELECT * FROM mytable WHERE phonenumber LIKE 'blah%'). This will handle sitations (although with the possibility of error) where the area code is not provided, but your database has the number with the area code.

  Brian Boatright : true. I did add something after the regex that returned the entire string if it was 10 digits or right(string,10) if it was longer. you last suggestion is a good one and something I will add. thanks! +1