Pfthb

OK, I've been trying to figure this one out for a while, but I can't seem to get it right. I need to delete the last occurrence of a non-consecutive, duplicate word/phrase from the END of a string. For example, I want

Love in My Antonia Love in

and

Love in My Antonia Love

to become

Love in My Antonia

I have tried countless patterns without success. The closest I have gotten to success is this:

 preg_replace('/b(w2,)b(?=.*?\1)W*/', ''

That deletes the first occurrence (rather than the last), rendering:

 in My Antonia Love (ORIGINAL: "Love in My Antonia Love")

and

 My Antonia Love in (ORIGINAL: "Love in My Antonia Love in")

Please help! :)

UPDATE (Nov. 9, 2:00 PM, PST): I should have clarified that—if at all possible—I would like the solution to remain in the simple, 1-line, compact format that I showed in my example:

preg_replace('/b(w2,)b(?=.*?\1)W*/', ''

My example already works almost perfectly, except that it deletes the first match instead of the last. I was hoping that someone could modestly manipulate my existing code so that it deletes the last match (at the end of the string) instead of the first occurrence. Is that more complicated than I thought?

Previously, I came up with a version that finds two consecutive, duplicate words/phrases anywhere in a string and replaces them with one:

preg_replace('~b([S w]3,)Kb(?:s*1)+~', '', 

This makes "pizza pizza" become "pizza" and "I walked to the store I walked to the store" become "I walked to the store." That's great, and I have already incorporated that solution. Now, secondarily, I also need "Pizza is the best pizza" to become "Pizza is the best." Similarly, "Sheep dogs are awesome pets dogs are" should become "Sheep dogs are awesome pets." So, basically, it does not matter where in the string the first occurrence appears; all that matters is that the occurrence at the END OF THE STRING gets deleted. I hope this brings more clarity.

Here you go:

$s = preg_replace('/^b([w ]+)(.*?)b(1)$/i', '\1\2', $s);

Test:

$s = "Love in My Antonia Love in";
$s1 = "Love in My Antonia Love";
$s2 = "Love in My Antonia Love Not On End";

echo "Original:n$sn";
echo preg_replace('/^b([w ]+)(.*?)b(1)$/i', '\1\2', $s);
echo "n";
echo "Original:n$s1n";
echo preg_replace('/^b([w ]+)(.*?)b(1)$/i', '\1\2', $s1);
echo "n";
echo "Original:n$s2n";
echo preg_replace('/^b([w ]+)(.*?)b(1)$/i', '\1\2', $s2);

Output:

ZC-MGMT-04:~ jv$ php -q c.php
Original:
Love in My Antonia Love in
Love in My Antonia
Original:
Love in My Antonia Love
Love in My Antonia
Original:
Love in My Antonia Love Not On End
Love in My Antonia Love Not On End

====

UPDATE:

Jason suggests slight update to address ' at the end of the word:

preg_replace('/^b([w ]+)(.*?)bb(1)('s)*b$/i', '\1\2')

You could approach this without regular expressions, by splitting the sentence into component words and then manually checking the last two words:

$input = "Love in My Antonia Love in";
$words = preg_split("/s+/", $input);
$last = $words[count($words)-1];
$pattern = "/^(?=.*b" . $last . "b.*b" . $last . "b).*/";
if ($words[count($words) - 1] != $words[count($words) - 2] &&
 preg_match($pattern, $input, $match)) 
 array_pop($words);

$output = implode(" ", $words);
echo $input . "n" . $output;

Love in My Antonia Love in
Love in My Antonia Love

You need to find the longest duplicate substring first then remove it from the end of subject string. This could be done using a preg_match_all for doing a case-insensitive search then a preg_replace for omitting them:

$str = 'Love in My Antonia Love in';
preg_match_all('~(bw++(?> w++)*)(?=.*?b1)~i', $str, $matches);
$array = array_unique(array_map('strtolower', $matches[1]));
foreach ($array as $value) 
 $str = preg_replace("~^.*K(?<!s)s*b$value~i", '', $str);

echo trim($str); // Love in My Antonia

See live demo here