Split string on non-alphanumeric characters and on positions between digits and non-digits

I'm trying to split a string by non-开发者_JS百科alphanumeric delimiting characters AND between alternations of digits and non-digits. The end result should be a flat array of consisting of alphabetic strings and numeric strings.

I'm working in PHP, and would like to use REGEX.

Examples:

• ES-3810/24MX should become ['ES', '3810', '24', 'MX']
• CISCO1538M should become ['CISCO' , '1538', 'M']

The input file sequence can be indifferently DIGITS or ALPHA.

The separators can be non-ALPHA and non-DIGIT chars, as well as a change between a DIGIT sequence to an APLHA sequence, and vice versa.

---

The command to match all occurrances of a regex is preg_match_all() which outputs a multidimensional array of results. The regex is very simple... any digit ([0-9]) one or more times (+) or (|) any letter ([A-z]) one or more times (+). Note the capital A and lowercase z to include all upper and lowercase letters.

The textarea and php tags are inluded for convenience, so you can drop into your php file and see the results.

<textarea style="width:400px; height:400px;">
<?php

foreach( array(
        "ES-3810/24MX",
        "CISCO1538M",
        "123ABC-ThatsHowEasy"
    ) as $string ){

    // get all matches into an array
    preg_match_all("/[0-9]+|[[:upper:][:lower:]]+/",$string,$matches);

    // it is the 0th match that you are interested in...
    print_r( $matches[0] );

}

?>
</textarea>

Which outputs in the textarea:

Array
(
    [0] => ES
    [1] => 3810
    [2] => 24
    [3] => MX
)
Array
(
    [0] => CISCO
    [1] => 1538
    [2] => M
)
Array
(
    [0] => 123
    [1] => ABC
    [2] => ThatsHowEasy
)

---

$str = "ES-3810/24MX35 123 TEST 34/TEST";
$str = preg_replace(array("#[^A-Z0-9]+#i","#\s+#","#([A-Z])([0-9])#i","#([0-9])([A-Z])#i"),array(" "," ","$1 $2","$1 $2"),$str);
echo $str;
$data = explode(" ",$str);
print_r($data);

I could not think on a more 'cleaner' way.

---

The most direct preg_ function to produce the desired flat output array is preg_split().

Because it doesn't matter what combination of alphanumeric characters are on either side of a sequence of non-alphanumeric characters, you can greedily split on non-alphanumeric substrings without "looking around".

After that preliminary obstacle is dealt with, then split on the zero-length positions between a digit and a non-digit OR between a non-digit and a digit.

/             #starting delimiter
[^a-z\d]+     #match one or more non-alphanumeric characters
|             #OR
\d\K(?=\D)    #match a number, then forget it, then lookahead for a non-number
|             #OR
\D\K(?=\d)    #match a non-number, then forget it, then lookahead for a number
/             #ending delimiter
i             #case-insensitive flag

Code: (Demo)

var_export(
    preg_split('/[^a-z\d]+|\d\K(?=\D)|\D\K(?=\d)/i', $string, 0, PREG_SPLIT_NO_EMPTY)
);

preg_match_all() isn't a silly technique, but it doesn't return the array, it returns the number of matches and generates a reference variable containing a two dimensional array of which the first element needs to be accessed. Admittedly, the pattern is shorter and easier to follow. (Demo)

var_export(
    preg_match_all('/[a-z]+|\d+/i', $string, $m) ? $m[0] : []
);