PHP preg_split or preg_match sentences but keep punctuation in Array

Question

I am looking to breakdown a paragraph into sentences and then into 'exploded' strings but need to keep the punctuations as elements of the array.

Example text:

$meta = 'I am looking to break this paragraph into chunks. 
         I have researched, tried and tested various combinations; however, I cannot
         seem to make it work. Would anyone help me figure this out? 
         I thank you in advance...'

The desired output would be:

Array ( [0] =>
           Array ( [0] => I [1] => am [2] => looking [3] => to [4] => break [5] => [6] => this [7] => paragraph [8] => into [9] => chunks [10] => . )
        [1] =>    
           Array ( [0] => I [2] => have [3] => researched [4] => , [5] => tried [......
          ......] [5] => figure [6] => this [7] => out [8] => ? )
        [3] =>
          Array ( [0] => I [1] => thank [2] => you [3] => in [4] => advance [5] => ... )
      )

I have tried using:

$s = preg_split('/\s*[!?.]\s*/u', $meta, -1, PREG_SPLIT_NO_EMPTY);

to separate out the sentences but whilst this works, the punctuation disappears.

I would really appreciate help with building this two level array with the punctuation

Answer 1

You could do what you want using preg_match:

$meta = 'I am looking to break this paragraph into chunks. 
         I have researched, tried and tested various combinations; however, I cannot
         seem to make it work. Would anyone help me figure this out? 
         I thank you in advance...';

preg_match_all('/(\w+|[.;?,]+)/', $meta, $m);
print_r($m);

Explanation:

/           : regex delimiter
  (         : begin group 1
    \w+     : 1 or more aphanumeric character <=> [a-zA-Z0-9_]
    |       : OR
    [.;?,]+ : 1 or more punctuation
  )         : end of group 1
/           : regex delimiter

This will match and store in group 1 evry word an every group of punctuiation character.

If you want to be unicode compatible, you could use \p{L} for any letter and \p{P} for punctuation:

/(\p{L}+|\p{P}+)/

Output:

Array
(
    [0] => Array
        (
            [0] => I
            [1] => am
            [2] => looking
            [3] => to
            [4] => break
            [5] => this
            [6] => paragraph
            [7] => into
            [8] => chunks
            [9] => .
            [10] => I
            [11] => have
            [12] => researched
            [13] => ,
            [14] => tried
            [15] => and
            [16] => tested
            [17] => various
            [18] => combinations
            [19] => ;
            [20] => however
            [21] => ,
            [22] => I
            [23] => cannot
            [24] => seem
            [25] => to
            [26] => make
            [27] => it
            [28] => work
            [29] => .
            [30] => Would
            [31] => anyone
            [32] => help
            [33] => me
            [34] => figure
            [35] => this
            [36] => out
            [37] => ?
            [38] => I
            [39] => thank
            [40] => you
            [41] => in
            [42] => advance
            [43] => ...
        )

    [1] => Array
        (
            [0] => I
            [1] => am
            [2] => looking
            [3] => to
            [4] => break
            [5] => this
            [6] => paragraph
            [7] => into
            [8] => chunks
            [9] => .
            [10] => I
            [11] => have
            [12] => researched
            [13] => ,
            [14] => tried
            [15] => and
            [16] => tested
            [17] => various
            [18] => combinations
            [19] => ;
            [20] => however
            [21] => ,
            [22] => I
            [23] => cannot
            [24] => seem
            [25] => to
            [26] => make
            [27] => it
            [28] => work
            [29] => .
            [30] => Would
            [31] => anyone
            [32] => help
            [33] => me
            [34] => figure
            [35] => this
            [36] => out
            [37] => ?
            [38] => I
            [39] => thank
            [40] => you
            [41] => in
            [42] => advance
            [43] => ...
        )

)