Is there any faster way instead of using preg_match in the following code?

Question

this code finds if there is the string 2004 in <date_iso></date_iso> and if it is so, I echo some data from that specific element that the search string was found.

I was wondering if this is the best/fastest approach because my main concern is speed and the XML file is huge. Thank you for your ideas.

this is a sample of the XML

<entry ID="4406">
 <id>4406</id>
 <title>Book Look Back at 2002</title>
 <link>http://www.sebastian-bergmann.de/blog/archives/33_Book_Look_Back_at_2002.html</link>
 <description></description>
 <content_encoded></content_encoded>
 <dc_date>20.1.2003, 07:11</dc_date>
 <date_iso>2003年01月20日T07:11</date_iso>
 <blog_link/>
 <blog_title/>
</entry>

this is the code

<?php
$books = simplexml_load_file('planet.xml');
$search = '2004';
foreach ($books->entry as $entry) {
 if (preg_match('/' . preg_quote($search) . '/i', $entry->date_iso)) {
 echo $entry->dc_date;
 }
}
?>

Answer 1

First of all, put up a timer so you know if things get better.

You're repeating '/' . preg_quote($search) . '/i'for each book. You should create the search string only once or else you are wasting time:

<?php
$books = simplexml_load_file('planet.xml');
$search = '2004';
$regex = '/' . preg_quote($search) . '/i';
foreach ($books->entry as $entry) {
 if (preg_match($regex, $entry->date_iso)) {
 echo $entry->dc_date;
 }
}
?>

If you are only looking for 2004 or similar you might analyze if simpler functions would be faster e.g. strpos.

Also the /i modifier might be unnecessary.

Answer 2

Be aware, that if you use strpos you should use the !== operator to check for an occurance of your haystack. If 2004 is at position 0 the != will evaluate true.

<?php
$books = simplexml_load_file('planet.xml');
$search = '2004';
foreach ($books->entry as $entry) {
 if (strpos($entry->date_iso, $search) !== false) {
 echo $entry->dc_date;
 }
}
?>

Answer 3

This might not exactly be an answer you're looking for, but I would imagine that the function simplexml_load_file is somewhat expensive on a large XML file since it creates a lot of objects containing information about the elements, attributes, values, contents, etc.

For that reason I would try finding matching elements with preg_match first (match entire entry tags that contain 2004 in their respective date_iso tags), then either load just these matches with simplexml_load_file to extract the desired information, or (possibly even better) use preg_match to do the same thing.

• takeshin: simplexml is simple, preg_match (in a loop!) is way more expensive.
• newenglander: Are you sure about that? I would think it would depend on the XML and what exactly you're looking for in it.