I always have trouble with regex, I basically have a url, for example:
http://somedomain.com/something_here/bla/bla/bla/bla.jpg
What I need is a preg_replace() to replace the something_here with an empty string, and leave everything else in tact.
I have tried the following and it replaces the wrong parts:
$image[0] = preg_replace('/http:\/\/(.*)\/(.*)\/wp-content\/uploads\/(.*)/','$2' . '',$image[0]);
This ends up leaving only the part I want to replace, rather than actually replacing it!
The following code is based on the description you provided:
$url = 'http://somedomain.com/something_here/bla/bla/bla/bla.jpg';
$output = preg_replace('#^(https?://[^/]+/)[^/]+/(.*)$#', '$1$2', $url);
echo $output; // http://somedomain.com/bla/bla/bla/bla.jpg
Explanation:
^ : match begin of line( : start matching group 1
https?:// : match http or https protocol[^/]+ : match anything except / one or more times/ : match /) : end matching group 1[^/]+ : match anything except / one or more times
-/ : match /( : start matching group 2
.* : match anything zero or more times (greedy)) : end matching group 2$ : match end of lineYou could do this:
$image[0] = preg_replace('!^(http://[^/]*)/[^/]*!', '$1', $image[0]);
Or you might consider just splitting the string to work on its individual components:
$parts = explode('/', $image[0]);
unset($parts[3]);
$image[0] = implode('/', $parts);
You could do this with a simple string replace:
$image[0] = str_replace('/wp-content/uploads/', '/', $image[0]);
Or if you want to use a regular expression:
$image[0] = preg_replace('~(http://.*?)/wp-content/uploads/(.*)~', '$1/$2', $image[0]);
/ you need to put value from a variable that you do not control, then you need to escape the $ and \number characters, and this is quite a difficult task. Use preg_replace_callback, an example is in my answer to the question.Unreliable usage:
In preg_replace('/(:\/\/.*?\/).*?\/()/', '$1$2', $string): if instead of empty string you need to put value from a variable that you do not control, then you need to escape the $ and \number characters, and this is quite a difficult task.
Reliable usage:
For difficult regular expressions where you need to replace part of a string, it can be very useful to do so (and most importantly, you don't need to escape $newValue for the characters $ and \number):
$newValue = '';
$newString = preg_replace_callback(
'/(:\/\/.*?\/).*?\/()/',
function($match) use ($newValue) { return $match[1].$newValue.$match[2]; },
$string
);
parse_url()for this and do the replacement on the path portion.