PHP remove next two words after a specific word

Question

how can I remove the next two words after a specific word with preg_replace in PHP? For example: String: Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam. Specific word: ipsum New String: Lorem ipsum amet, consetetur sadipscing elitr, sed diam.

Thats my current code:

$txt = "Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam.
Specific word: ipsum";
$str= preg_replace('/\W\w+\s*(\W*)$/', '$1', $txt);
echo $str;

But is is just removing the last word of a string.

Thank you With best regards

Answer 1

You can use (?<=ipsum)(?: \w+){2}, but if you want to include punctuation marks use (?<=ipsum)(?: [A-Za-z,.!]+){2}.

function remove2w($anchor, $text, $number = 2) {
    return preg_replace(sprintf('/(?<=%s)(?: \w+){%s}/', $anchor, $number), '', $text);
}

Output:

remove2w('ipsum', 'Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam.')
>>> Lorem ipsum amet, consetetur sadipscing elitr, sed diam.

Answer 2

preg_replace() offers quite some flexibility:

<?php
$needle = "ipsum";
$haystack = "Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam. ";
$pattern = sprintf('|(%s)\s+\w+\s+\w+|', $needle);
var_dump(preg_replace($pattern, '$1', $haystack));

The output obviously is:

string(57) "Lorem ipsum amet, consetetur sadipscing elitr, sed diam. "

Answer 3

Another way using explode(). You can split the string by , then array_search() for your $word which will give you the index in the array, then simply unset() the next 2 elements:

<?php
$txt = "Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam.";
$word = "ipsum";
$txtArr = explode(" ", $txt);
$i = array_search($word, $txtArr);
unset($txtArr[$i + 2]);
unset($txtArr[$i + 1]);
var_dump(implode(" ", $txtArr));

Result

Lorem ipsum amet, consetetur sadipscing elitr, sed diam

Demo

Note: you'll need to do some error handling in case the $word is not found

Answer 4

There are many considerations to bake into this task.

1. Do you need word boundaries for your target substring? Without them you may perform unintentional matching; but only you can decide this point for your project.
2. Do you need case-insensitive matching? I'll guess: yes.
3. What happens if the target substring is the second last or last word in the string? Do you want to allow one or zero words to be omitted? I'll guess: yes.
4. You need to consider/include punctuation, right? I'll guess: yes.
5. Might your target substring contain regex-sensitive characters? If so, preg_quote() is recommended. I'll guess: no, but you can call preg_quote() on the needle before injecting it into the pattern if you are unsure.

Here is a complete battery of needles: (Demo)

$txt = 'Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam.';
$needles = str_word_count(strtolower($txt),1);
foreach($needles as $needle){
    echo "($needle) => ",preg_replace('~\b'.$needle.'\b\S*\K(?:\s\S+){0,2}~i','',$txt),"\n";  // use '(($0))'' as replacement to see the substring that is removed
}

Output:

(lorem) => Lorem sit amet, consetetur sadipscing elitr, sed diam.
(ipsum) => Lorem ipsum amet, consetetur sadipscing elitr, sed diam.
(dolor) => Lorem ipsum dolor consetetur sadipscing elitr, sed diam.
(sit) => Lorem ipsum dolor sit sadipscing elitr, sed diam.
(amet) => Lorem ipsum dolor sit amet, elitr, sed diam.
(consetetur) => Lorem ipsum dolor sit amet, consetetur sed diam.
(sadipscing) => Lorem ipsum dolor sit amet, consetetur sadipscing diam.
(elitr) => Lorem ipsum dolor sit amet, consetetur sadipscing elitr,
(sed) => Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed
(diam) => Lorem ipsum dolor sit amet, consetetur sadipscing elitr, sed diam.

Breakdown:

~                #pattern delimiter
\b'.$needle.'\b  #match needle as a whole word
\S*              #match zero or more trailing character as long as first character is a non-word character.  This may be replaced with [[:punct:]]+ if more desirable/accurate
\K               #restart fullstring match
(?:\s\S+){0,2}   #match zero, one or two sequences of: a whitespace character followed by one or more non-whitespace characters
~                #pattern delimiter
i                #case-insensitive pattern modifier