Regular Expression to reformat a US phone number in Javascript

I'm looking to reformat (replace, not validate - there are many references for validating) a phone number for display in Javascript. Here's an example of some of the data:

• 123 4567890
• (123) 456-7890
• (123)456-7890
• 123 456 7890
• 123.456.7890
• (blank/null)
• 1234567890

Is there an easy way to use a regular expression to do this? I'm looking for the best way to do this. Is there a better way?

I want to reformat the number to the following: (123) 456-7890

Assuming you want the format "(123) 456-7890":

function formatPhoneNumber(phoneNumberString) {
  var cleaned = ('' + phoneNumberString).replace(/\D/g, '');
  var match = cleaned.match(/^(\d{3})(\d{3})(\d{4})$/);
  if (match) {
    return '(' + match[1] + ') ' + match[2] + '-' + match[3];
  }
  return null;
}

Here's a version that allows the optional +1 international code:

function formatPhoneNumber(phoneNumberString) {
  var cleaned = ('' + phoneNumberString).replace(/\D/g, '');
  var match = cleaned.match(/^(1|)?(\d{3})(\d{3})(\d{4})$/);
  if (match) {
    var intlCode = (match[1] ? '+1 ' : '');
    return [intlCode, '(', match[2], ') ', match[3], '-', match[4]].join('');
  }
  return null;
}
formatPhoneNumber('+12345678900') // => "+1 (234) 567-8900"
formatPhoneNumber('2345678900')   // => "(234) 567-8900"

Possible solution:

function normalize(phone) {
    //normalize string and remove all unnecessary characters
    phone = phone.replace(/[^\d]/g, "");

    //check if number length equals to 10
    if (phone.length == 10) {
        //reformat and return phone number
        return phone.replace(/(\d{3})(\d{3})(\d{4})/, "($1) $2-$3");
    }

    return null;
}

var phone = '(123)4567890';
phone = normalize(phone); //(123) 456-7890

var x = '301.474.4062';
    
x = x.replace(/\D+/g, '')
     .replace(/(\d{3})(\d{3})(\d{4})/, '($1) $2-$3');

alert(x);

This answer borrows from maerics' answer. It differs primarily in that it accepts partially entered phone numbers and formats the parts that have been entered.

phone = value.replace(/\D/g, '');
const match = phone.match(/^(\d{1,3})(\d{0,3})(\d{0,4})$/);
if (match) {
  phone = `${match[1]}${match[2] ? ' ' : ''}${match[2]}${match[3] ? '-' : ''}${match[3]}`;
}
return phone

I'm using this function to format US numbers.

function formatUsPhone(phone) {

    var phoneTest = new RegExp(/^((\+1)|1)? ?\(?(\d{3})\)?[ .-]?(\d{3})[ .-]?(\d{4})( ?(ext\.? ?|x)(\d*))?$/);

    phone = phone.trim();
    var results = phoneTest.exec(phone);
    if (results !== null && results.length > 8) {

        return "(" + results[3] + ") " + results[4] + "-" + results[5] + (typeof results[8] !== "undefined" ? " x" + results[8] : "");
    
    }
    else {
         return phone;
    }
}

It accepts almost all imaginable ways of writing a US phone number. The result is formatted to a standard form of (987) 654-3210 x123

thinking backwards

Take the last digits only (up to 10) ignoring first "1".

function formatUSNumber(entry = '') {
  const match = entry
    .replace(/\D+/g, '').replace(/^1/, '')
    .match(/([^\d]*\d[^\d]*){1,10}$/)[0]
  const part1 = match.length > 2 ? `(${match.substring(0,3)})` : match
  const part2 = match.length > 3 ? ` ${match.substring(3, 6)}` : ''
  const part3 = match.length > 6 ? `-${match.substring(6, 10)}` : ''    
  return `${part1}${part2}${part3}`
}

example input / output as you type

formatUSNumber('+1333')
// (333)

formatUSNumber('333')
// (333)

formatUSNumber('333444')
// (333) 444

formatUSNumber('3334445555')
// (333) 444-5555

2021

libphonenumber-js

Example

import parsePhoneNumber from 'libphonenumber-js'

const phoneNumber = parsePhoneNumber('+12133734253')

phoneNumber.formatInternational() === '+1 213 373 4253'
phoneNumber.formatNational() === '(213) 373-4253'
phoneNumber.getURI() === 'tel:+12133734253'

I've extended David Baucum's answer to include support for extensions up to 4 digits in length. It also includes the parentheses requested in the original question. This formatting will work as you type in the field.

phone = phone.replace(/\D/g, '');
const match = phone.match(/^(\d{1,3})(\d{0,3})(\d{0,4})(\d{0,4})$/);
if (match) {
	phone = `(${match[1]}${match[2] ? ') ' : ''}${match[2]}${match[3] ? '-' : ''}${match[3]}${match[4] ? ' x' : ''}${match[4]}`;
}
return phone;

Based on David Baucum's answer - here is a version that trys to improve auto-replacement "as you type" for example in a React onChange event handler:

function formatPhoneNumber(phoneNumber) {
  const cleanNum = phoneNumber.toString().replace(/\D/g, '');
  const match = cleanNum.match(/^(\d{3})(\d{0,3})(\d{0,4})$/);
  if (match) {
    return '(' + match[1] + ') ' + (match[2] ? match[2] + "-" : "") + match[3];
  }
  return cleanNum;
}

//...

onChange={e => setPhoneNum(formatPhoneNumber(e.target.value))}

It will insert (###) as soon as there are 3 numbers and then it will keep following the RegEx until it looks like this (###) ###-####

var numbers = "(123) 456-7890".replace(/[^\d]/g, ""); //This strips all characters that aren't digits
if (numbers.length != 10) //wrong format
    //handle error
var phone = "(" + numbers.substr(0, 3) + ") " + numbers.substr(3, 3) + "-" + numbers.substr(6); //Create format with substrings

Here is one that will accept both phone numbers and phone numbers with extensions.

function phoneNumber(tel) {
var toString = String(tel),
    phoneNumber = toString.replace(/[^0-9]/g, ""),
    countArrayStr = phoneNumber.split(""),
    numberVar = countArrayStr.length,
    closeStr = countArrayStr.join("");
if (numberVar == 10) {
    var phone = closeStr.replace(/(\d{3})(\d{3})(\d{4})/, "$1.$2.$3"); // Change number symbols here for numbers 10 digits in length. Just change the periods to what ever is needed.
} else if (numberVar > 10) {
    var howMany = closeStr.length,
        subtract = (10 - howMany),
        phoneBeginning = closeStr.slice(0, subtract),
        phoneExtention = closeStr.slice(subtract),
        disX = "x", // Change the extension symbol here
        phoneBeginningReplace = phoneBeginning.replace(/(\d{3})(\d{3})(\d{4})/, "$1.$2.$3"), // Change number symbols here for numbers greater than 10 digits in length. Just change the periods and to what ever is needed. 
        array = [phoneBeginningReplace, disX, phoneExtention],
        afterarray = array.splice(1, 0, " "),
        phone = array.join("");

} else {
    var phone = "invalid number US number";
}
return phone;
}

phoneNumber("1234567891"); // Your phone number here

Almost all of these have issues when the user tries to backspace over the delimiters, particularly from the middle of the string.

Here's a jquery solution that handles that, and also makes sure the cursor stays in the right place as you edit:

//format text input as phone number (nnn) nnn-nnnn
$('.myPhoneField').on('input', function (e){
    var $phoneField = e.target;
    var cursorPosition = $phoneField.selectionStart;
    var numericString = $phoneField.value.replace(/\D/g, '').substring(0, 10);

    // let user backspace over the '-'
    if (cursorPosition === 9 && numericString.length > 6) return;

    // let user backspace over the ') '
    if (cursorPosition === 5 && numericString.length > 3) return;
    if (cursorPosition === 4 && numericString.length > 3) return;

    var match = numericString.match(/^(\d{1,3})(\d{0,3})(\d{0,4})$/);
    if (match) {
        var newVal = '(' + match[1];
        newVal += match[2] ? ') ' + match[2] : '';
        newVal += match[3] ? '-' + match[3] : '';

        // to help us put the cursor back in the right place
        var delta = newVal.length - Math.min($phoneField.value.length, 14);      
        $phoneField.value = newVal;
        $phoneField.selectionEnd = cursorPosition + delta;
    } else {
        $phoneField.value = '';        
    }
})

For US Phone Numbers

/^\(?(\d{3})\)?[- ]?(\d{3})[- ]?(\d{4})$/

Let’s divide this regular expression in smaller fragments to make is easy to understand.

• /^\(?: Means that the phone number may begin with an optional (.
• (\d{3}): After the optional ( there must be 3 numeric digits. If the phone number does not have a (, it must start with 3 digits. E.g. (308 or 308.
• \)?: Means that the phone number can have an optional ) after first 3 digits.
• [- ]?: Next the phone number can have an optional hyphen (-) after ) if present or after first 3 digits.
• (\d{3}): Then there must be 3 more numeric digits. E.g (308)-135 or 308-135 or 308135
• [- ]?: After the second set of 3 digits the phone number can have another optional hyphen (-). E.g (308)-135- or 308-135- or 308135-
• (\d{4})$/: Finally, the phone number must end with four digits. E.g (308)-135-7895 or 308-135-7895 or 308135-7895 or 3081357895.

Reference :

http://www.zparacha.com/phone_number_regex/

You can use this functions to check valid phone numbers and normalize them:

let formatPhone = (dirtyNumber) => {
 return dirtyNumber.replace(/\D+/g, '').replace(/(\d{3})(\d{3})(\d{4})/, '($1) $2-$3');
}

let isPhone = (phone) => {
   //normalize string and remove all unnecessary characters
   phone = phone.replace(/\D+/g, '');
   return phone.length == 10? true : false;
}

The solutions above are superior, especially if using Java, and encountering more numbers with more than 10 digits such as the international code prefix or additional extension numbers. This solution is basic (I'm a beginner in the regex world) and designed with US Phone numbers in mind and is only useful for strings with just 10 numbers with perhaps some formatting characters, or perhaps no formatting characters at all (just 10 numbers). As such I would recomend this solution only for semi-automatic applications. I Personally prefer to store numbers as just 10 numbers without formatting characters, but also want to be able to convert or clean phone numbers to the standard format normal people and apps/phones will recognize instantly at will.

I came across this post looking for something I could use with a text cleaner app that has PCRE Regex capabilities (but no java functions). I will post this here for people who could use a simple pure Regex solution that could work in a variety of text editors, cleaners, expanders, or even some clipboard managers. I personally use Sublime and TextSoap. This solution was made for Text Soap as it lives in the menu bar and provides a drop-down menu where you can trigger text manipulation actions on what is selected by the cursor or what's in the clipboard.

My approach is essentially two substitution/search and replace regexes. Each substitution search and replace involves two regexes, one for search and one for replace.

Substitution/ Search & Replace #1

• The first substitution/ search & replace strips non-numeric numbers from an otherwise 10-digit number to a 10-digit string.

First Substitution/ Search Regex: \D

• This search string matches all characters that is not a digit.

First Substitution/ Replace Regex: "" (nothing, not even a space)

• Leave the substitute field completely blank, no white space should exist including spaces. This will result in all matched non-digit characters being deleted. You should have gone in with 10 digits + formatting characters prior this operation and come out with 10 digits sans formatting characters.

Substitution/ Search & Replace #2

• The second substitution/search and replace search part of the operation captures groups for area code $1, a capture group for the second set of three numbers $2, and the last capture group for the last set of four numbers $3. The regex for the substitute portion of the operation inserts US phone number formatting in between the captured group of digits.

Second Substitution/ Search Regex: (\d{3})(\d{3})(\d{4})

Second Substitution/ Replace Regex: \($1\) $2\-$3

• The backslash \ escapes the special characters (, ) , (<-whitespace), and - since we are inserting them between our captured numbers in capture groups $1, $2, & $3 for US phone number formatting purposes.

• In TextSoap I created a custom cleaner that includes the two substitution operation actions, so in practice it feels identical to executing a script. I'm sure this solution could be improved but I expect complexity to go up quite a bit. An improved version of this solution is welcomed as a learning experience if anyone wants to add to this.