Question

Using PHP, I am looking to extract an array from a string that contains a numbered list.

Example string:

The main points are: 1. This is point one. 2. This is point two. 3. This is point three.

would result in the following array:

[0] => 1. This is point one.
[1] => 2. This is point two.
[2] => 3. This is point three.

The format of the string can vary – e.g.:

1. This is point one, 2. This is point two, 3. This is point three.
1) This is point one  2) This is point two 3) This is point three
1 This is point one. 2 This is point two. 3 This is point three.

I have started using preg_match_all with the following pattern:

!((\d+)(\s+)?(\.?)(\)?)(-?)(\s+?)(\w+))!

but I am unsure as how to match rest of string/up to the next match.

Example available at RegExr

Answer 1

If your input follows your example input, as in each “point” doesn’t contain a number itself, you could use the following regex:

\d+[^\d]*

In PHP, you could use preg_match_all() to capture everything:

$text = 'The main points are: 1. This is point one. 2. This is point two. 3. This is point three.';

$matches = array();
preg_match_all('/(\d+[^\d]*)/', $text, $matches);

print_r($matches[1]);

This will result in:

Array
(
    [0] => 1. This is point one.
    [1] => 2. This is point two.
    [2] => 3. This is point three.
)

Again though, if there are any numbers/digits in the actual points themselves – this won’t work.

If you want actual numbers to appear in each point, you’ll need to define an actual “anchor” or “end” of each point, such as a period. If you can state that a . will appear only at the end of the point (ignoring the potential one that follows the leading-digit), you could use the following regex:

\d+[.)\s][^.]*\.

It can be dropped into the preg_match_all() from above just as easily:

preg_match_all('/(\d+[.)\s][^.]*\.)/', $text, $matches);

Regex explained:

\d+        # leading number
[.)\s]     # followed by a `.`, `)`, or whitespace
[^.]*      # any non-`.` character(s)
\.         # ending `.`

The caveat with the second regex is that a . may only appear at the end of each point (and following the leading digit). However, I think that this rule may be easier to follow than the “no numbers in the point” rule – it all depends on your actual input though.