Getting part of a string using REGEX - php

Question

I have an Amazon link:
http://www.amazon.com/Pampers-Softcare-Fresh-Wipes-Count/dp/B007KXO998/ref=pd_zg_rss_ts_165796011_165796011_7?ie=UTF8&tag=elson06-20
I'm trying to get the product ID B007FHX9OK that is after dp/ and before ?ref=pd_zg_rss_ts_165796011_165796011_7
I want to get that using a regex or anything that can extract it.
The link of the url is static, it will not changed.

Answer 1

$string = 'http://www.amazon.com/iOttie-Windshield-INCREDIBLE-BlackBerry-Revolution/dp/B007FHX9OK?SubscriptionId=AKIAJJPPYQPVMQLOYLKQ&tag=elson06-20&linkCode=sp1&camp=2025&creative=165953&creativeASIN=B007FHX9OK';
//$string = 'http://www.amazon.com/Pampers-Softcare-Fresh-Wipes-Count/dp/B007KXO998/ref=pd_zg_rss_ts_165796011_165796011_7?ie=UTF8&tag=elson06-20';
$pid = basename((false !== strpos($string, '/ref='))
? pathinfo($string, PATHINFO_DIRNAME)
: parse_url($string, PHP_URL_PATH));
echo $pid; // Outputs B007KXO998 or B007FHX9OK, will work for both types of URLs
You don't need a regex, PHP has built-in functions to parse URLs.

Answer 2

Will the URLs always be in this exact format, or will it be expected to match any Amazon URL?
If the format will always be like this, then you can use #cryptic's answer. Otherwise, it would be more flexible to use a pattern like |dp/([A-Z0-9]+)|i for the pattern.
This will match any alphanumeric string (case insensitive) directly following dp/ in the string. Well, the entire match will include the dp/ part, but the parenthetical portion is a sub-match which will match only the product id.
Edit: According to this page, Amazon's product IDs (ASINs) can be present in a wide variety of URLs, making them difficult to match, and my code above won't catch them all.
One way to try to catch these would be to use parse_url to extract the host and the path portions of the URL. From there, you can check the host portion against known Amazon domain names, and you could explode the path, and check each portion for an alphanumeric section which is ten characters long. Even then, the ASIN for books is the books ISBN, and there are 13-digit versions which Amazon might use in some cases (though I don't have evidence that they do).
Here is a very basic example that I haven't thoroughly tested:
$url = get_url_from_wherever();
$url_parts = parse_url($url);
$host = $url_parts['host'];
$path = explode('/', $url_parts['path']);
$amazon_hosts = array(
'amazon.com', // United States
'amazon.ca', // Canada
'amazon.cn', // China
'amazon.fr', // France
'amazon.it', // Italy
'amazon.de', // Germany
'amazon.es', // Spain
'amazon.co.jp', // Japan
'amazon.co.uk', // United Kingdom
'amzn.to' // URL Shortener
);
$amazon_hosts = array_map('preg_quote', $amazon_hosts);
$asin = FALSE; // initialize in case we don't find the ASIN
if (preg_match('/(^|\.)(' . implode($amazon_hosts, '|') . ')$/i', $host)) {
// valid host
foreach($path as $path_component) {
if (preg_match('/^[A-Z0-9]{10}$/i', $path_component)) {
// this is probably the ASIN, since the string is a 10-character alphanumeric
$asin = $path_component;
}
}
}
if ($asin) {
// process ASIN
} else {
// couldn't find an ASIN in this URL
}

Answer 3

Here's what I did, since I'm pretty sure that the link has always the same format:
$link = 'http://www.amazon.com/Pampers-Softcare-Fresh-Wipes-Count/dp/B007KXO998/ref=pd_zg_rss_ts_165796011_165796011_7?ie=UTF8&tag=elson06-20'
$link = parse_url($link);
$link = explode('/',$link['path']);
$link = $link[3];
echo $link; //B007KXO998