Using preg_replace_callback() to produce a running sequence - PHP

We will now take a look at an even more interesting example – doing a preg_replace using a regular expression with a running sequence.

Suppose you have an ordered list in HTML format:

list 1

<ol>

<li>item A</li>

<li>item B</li>

<li>item C</li>

</ol>

You would like to convert this into a plain text format in running sequence:

list 1

1. item A

2. item B

3. item C

Below is the code using preg_replace_callback():

<?php

$str = "list 1

<ol>

<li>item A</li>

<li>item B</li>

<li>item C</li>

</ol>

";

echo preg_replace_callback('%<ol>(.*)</ol>%si', 'process_list', $str);

function process_list($matches) {

return preg_replace_callback("/<li>(.*)</li>/i", 'process_item', trim($matches[1]));

}

function process_item($matches) {

static $sno=0;

++$sno;

return "$sno. $matches[1]";

}

?>

Note the use of two preg_replace_callback() functions. The first one is used to grab all the items within the tags <ol> and </ol>. The second one is used to process each individual item, in which we append a running serial number in front of each item.

In case you have multiple lists such as:

list 1

<ol>

<li>item A</li>

<li>item B</li>

<li>item C</li>

</ol>

list 2

<ol>

<li>item D</li>

<li>item E</li>

<li>item F</li>

<li>item G</li>

</ol>

list 3

<ol>

<li>item H</li>

<li>item I</li>

<li>item J</li>

<li>item K</li>

<li>item L</li>

</ol>

you have to change the code slightly to accommodate this:

<?php

$str = "

list 1

<ol>

<li>item A</li>

<li>item B</li>

<li>item C</li>

</ol>

list 2

<ol>

<li>item D</li>

<li>item E</li>

<li>item F</li>

<li>item G</li>

</ol>

list 3

<ol>

<li>item H</li>

<li>item I</li>

<li>item J</li>

<li>item K</li>

<li>item L</li>

</ol>

";

echo preg_replace_callback('%<ol>(.*?)</ol>%si', 'process_list', $str);

function process_list($matches) {

global $sno;

$sno = 0;

return preg_replace_callback("/<li>(.*?)</li>/i", 'process_item', trim($matches[1]));

}

function process_item($matches) {

global $sno;

++$sno;

return "$sno. $matches[1]";

}

?>

Note the two key changes here. First, we have used the non-greedy quantifier ‘*?’ so that each list remains as one unit. Second, we need to think of some way to make the serial number restart from one for each new list. Here, for simplicity, I just use a global variable.