Extract phone numbers from string using regex?

Question

I need to extract some phone numbers from large strings in rails. These numbers will come in a variety of formats and could have multiple phone numbers in a single string.

Here is an example of the types of formats that occur:

022 1234567
021 123 2345
0271233211
021-233-9123
09 123 32112
021 2331231 or 021 321123123

What is the most efficient way to extract phone numbers like this that appear within a body of text?

UPDATE: Thanks for the answers. After testing some of them I realise that I should include more examples. Here are some more that don't appear in the list above:

622 32281
5754321
092213212
(09)1234321

This is possibly a duplicate of http://stackoverflow.com/questions/123559/a-comprehensive-regex-for-phone-number-validation — the Tin Man, Jul 21 '13 at 03:45
You can quickly test your regex matches against all your test samples at rubular.com. It has a nice visual feedback component showing which characters are getting matched. — HM1, Jul 21 '13 at 04:15

Andy G · Accepted Answer · 2013-07-21T00:24:39.140

6

I would keep it simple:

\d{2}[\s\d-]+

Two numbers, one or more of whitespace, numbers or a hyphen.

Require more characters with:

\d{2}[\s\d-]{5,}

(two numbers and 5 or more of whitespace, numbers of hyphens) which will reduce the number of mis-hits.

These will include an extra space following the phone-number, but the results could be trimmed.

Rather than trim, though, I would remove the hyphens and whitespace and count the number of digits leftover to recognise them as phone numbers.

If the phone numbers always start with a 0:

0\d[\s\d-]{5,}\d

this ends with a number, so drops the space at the end in the earlier examples.

Added following the further examples:

\b[\s()\d-]{6,}\d\b

edited Jul 21 '13 at 00:24

answered Jul 20 '13 at 23:45

Andy G

19,232
5
47
69

Thanks will give this a go. I have updated my question with some more examples. – startupsmith Jul 21 '13 at 00:17
The brackets and 7 digits throw out most of the responses so far, doh! – Andy G Jul 21 '13 at 00:20
That will pick up 2 digits followed by 5 spaces. I think the most important condition needs to be 7 digits. – pguardiario Jul 21 '13 at 10:58
Why would there be 2 digits followed by 5 spaces..? In any case, subsequently removing all the whitespace (and hyphens) would reveal that this is not a phone number. We could use \D to keep only the numbers. – Andy G Jul 21 '13 at 11:06
Why wouldn't there be? – pguardiario Jul 22 '13 at 04:06

score 4 · Answer 2 · edited May 23 '17 at 11:58

Here's how I'd go about it:

LOREM_IPSUM = "Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.".split
STRING = [
  '123 123 1234',
  LOREM_IPSUM.shift(1 + rand(4)),
  '123-123-1234',
  LOREM_IPSUM.shift(1 + rand(4)),
  '12 123 12345',
  LOREM_IPSUM.shift(1 + rand(4)),
  '123 1234567',
  LOREM_IPSUM.shift(1 + rand(4)),
  '123 123456789',
  LOREM_IPSUM.shift(1 + rand(4)),
  '123 12345',
  LOREM_IPSUM.shift(1 + rand(4)),
  '1234567',
  LOREM_IPSUM.shift(1 + rand(4)),
  '1234567890',
  LOREM_IPSUM.shift(1 + rand(4)),
  '123456789',
  LOREM_IPSUM.shift(1 + rand(4)),
  '(12)1234567',
].join(' ')

STRING # => "123 123 1234 Lorem ipsum dolor sit 123-123-1234 amet, consectetur adipisicing 12 123 12345 elit, sed do eiusmod 123 1234567 tempor 123 123456789 incididunt ut 123 12345 labore 1234567 et dolore magna aliqua. 1234567890 Ut enim ad minim 123456789 veniam, (12)1234567"
STRING.scan(/\d+.\d+.\d+/) # => ["123 123 1234", "123-123-1234", "12 123 12345", "123 1234567", "123 123456789", "123 12345", "1234567", "1234567890", "123456789", "12)1234567"]
STRING.scan(/\d+.\d+.\d+/).map{ |s| s.gsub(/\D+/, '') } # => ["1231231234", "1231231234", "1212312345", "1231234567", "123123456789", "12312345", "1234567", "1234567890", "123456789", "121234567"]

I removed a couple duplicate formats to simplify the test.

There are a lot of ways that a phone number can be formatted. "A comprehensive regex for phone number validation" is a good starting point for ideas. Based on the comment in the selected answer:

just strip all non-digit characters on input (except 'x')

I figure this is a reasonable starting pattern:

/\d+.\d+.\d+/

Using that with scan on the test string captures all the sample phone numbers above. Once you have them follow the next piece of advice in that answer:

[...] Then when you display, reformat to your hearts content.

Benjamin Bouchet · Answer 3 · 2013-07-20T23:47:48.373

1

I would use this

\b(\d{2}[\s|\-|\d]{2}\d{2}[\s|\d][\s|\-|\d]\d{2,5})\b

edited Jul 20 '13 at 23:47

answered Jul 20 '13 at 23:41

Benjamin Bouchet

12,971
2
41
73

Instead of surrounding with "not a digit" (which includes that character and requires at least one character of some other class), why not use \b (word boundary)? – Chip Camden Jul 20 '13 at 23:46
Because I was waiting for someone to notice :) :) – Benjamin Bouchet Jul 20 '13 at 23:48
Thanks for this. I have given some more examples that haven't been picked up by this regex. – startupsmith Jul 21 '13 at 00:17

oppure · Answer 4 · 2015-10-29T11:59:04.217

1

I've written this one ((\+\d+\s*|00\d+\s*|0\d+\s*)(\(\d+\)\s*|\d+\s*)?(\d{2,10}(\-|\/|\s)*){3,8})\b it works well as long as the number starts with a + a 0 or 00, this is required to avoid stripping other non phone groups of digits.

edited Oct 29 '15 at 11:59

answered Oct 29 '15 at 08:55

oppure

11
2

'((\+\s*\d+\s*|00\d+\s*|0\d+\s*)(\(\d+\)\s*|\d+\s*)?(\d{2,10}(\-|\/|\s)*){2,8})\b' just slightly modifyed to catch also number with a space after the + and with only 2 groups of digits (7 digits numbers) – oppure Nov 07 '15 at 12:08

score 0 · Answer 5 · answered Jul 20 '13 at 23:43

0

The general problem of recognizing phone numbers is pretty tricksy. But given your examples above, how about:

/\d{2,3}[\s-]?\d{3}[\s-]?\d{4,}/

two or three digits, optional space or dash, three digits, optional space or dash, four or more digits.

answered Jul 20 '13 at 23:43

Chip Camden

210
1
7

score 0 · Answer 6 · answered Jul 21 '13 at 11:05

0

I'm surprised to not see any 7's in anyone's answer. Here's one that will pick up all but the last one:

/(?=(?:\d[ -]*){7,})([\d -]*)/

Maybe you could strip out the ()'s first.

answered Jul 21 '13 at 11:05

pguardiario

53,827
19
119
159

Extract phone numbers from string using regex?

6 Answers6

Linked