preg_split is not breaking the strings in the array

Question

I’m not getting to use the function preg_split() correctly.

I’m trying to break a string array() via regex but it’s not happening.

$string = "<:termo.9:><:termo.10:><:termo.11:>";
$res = preg_split("/(<:)(nome|termo)(.)(\d+)(:>)/", $string,-1,PREG_SPLIT_DELIM_CAPTURE);

The value attributed to $res is:

        Array ( 
          [0] => 
          [1] => <: 
          [2] => termo 
          [3] => . 
          [4] => 9 
          [5] => :> 
          [6] => 
          [7] => <: 
          [8] => termo 
          [9] => . 
          [10] => 12 
          [11] => :> 
          [12] => 
          [13] => <: 
          [14] => termo 
          [15] => . 
          [16] => 10 
          [17] => :> 
          [18] => 
          [19] => <: 
          [20] => termo 
          [21] => . 
          [22] => 11 
          [23] => :> 
        )

And I’d actually like to:

Array (
           [0] => "<:termo.9:>"
           [1] => "<:termo.10:>"
           [2] => "<:termo.11:>"
       )

A palliative solution would be to use the following Pattern:

$res = preg_split("/(<:nome\.\d+:>|<:termo\.\d+:>)/", $string,-1,PREG_SPLIT_DELIM_CAPTURE);

And the $res sai:

Array ( 
          [0] => 
          [1] => <:termo.9:> 
          [2] => 
          [3] => <:termo.12:> 
          [4] => 
          [5] => <:termo.10:> 
          [6] => 
          [7] => <:termo.11:> 
          [8] => 
          [9] => <:termo.9:> 
          [10] => 
        )

But if you look closely, the conditions imposed are:

(<:nome\.\d+:>|<:termo\.\d+:>)

And the variations of nomenclatures that I can use can increase, like:

<:name.2:>, <:text.4:>,<:code.5:>,etc

and this way the code would not be "optimized", say so.

What could be wrong in Pattern for it to be improved?

Answer 1

The problem is not quite the pattern, the function preg_split, searching for the pattern to break the string into parts, that is to say it is similar to explode, however, in it you can define an explosion pattern that is not limited to the same sequence always.

Example

$string = "teste_de_explosão-de-string";
$res = preg_split('/_|-/', $string); // vai explodir por '_' e por '-'

In your case it would be a capture and not an explosion, and for that can be used the preg_match_all.

In your case

$string = "<:termo.9:><:termo.10:><:termo.11:>";
preg_match_all("/(<:(nome|termo)\.\d+:>)/", $string, $res);

To recover previously grouped values, use:

foreach($res[1] as $id=>$val){
    $arr[] = $val;  
}

print_r($arr);

Obs

To capture . literal in REGEX should escape the same \., for he represents anything.

Answer 2

$foo = "<:termo.9:><:termo.10:><:termo.11:><:nome.2:><:text.4:><:codigo.5:>";

preg_match_all("/<:[a-z]+\.[0-9]+:>/s", $foo, $bar);

print_r($bar);

Will return:

Array
(
    [0] => Array
        (
            [0] => <:termo.9:>
            [1] => <:termo.10:>
            [2] => <:termo.11:>
            [3] => <:nome.2:>
            [4] => <:text.4:>
            [5] => <:codigo.5:>
        )

)