正则表达式删除尾随字符

Question

I'm looking for a regular expression in Php that could transform incoming strings like this:

abaisser_negation_pronominal_question     => abaisser_n_p_q
abaisser_pronominal_question              => abaisser_p_q
abaisser_negation_question                => abaisser_n_q
abaisser_negation_pronominal              => abaisser_n_p
abaisser_negation_voix_passive_pronominal => abaisser_n_v_p_p
abaisser                                  => abaisser

With the Php code close to something like:

$line=preg_replace("/<h3>/im", "", $line);

How would you do?

Answer 1

You can use:

$input = preg_replace('/(_[A-Za-z])[^_\n]*/', '$1', $input);

RegEx Demo

Explanation:

• This regex searches for (_[A-Za-z])[^_\\n]* which means underscore followed by single letter and then match before a newline or underscore
• It capture first part (_[A-Za-z]) in a backreference $1
• Replacement is $1 leaving underscore and first letter in the replacement string

Answer 2

$line = preg_replace("/_([a-z])([a-z]*)/i", "_$1", $line);

Answer 3

You can use regex

$input = preg_replace('/_(.)[^\n_]+/', '_$1', $input);

DEMO

What it does is capture the character after _ and match till \\n or _ is encountered and replaced with the _$1 which means _ plus the character captured.

Answer 4

You could use \\K or positive lookbehind.

$input = preg_replace('~_.\K[^_\n]*~', '', $input);

Pattern _. in the above regex would match an _ and also the character following the underscore. \\K discards the previously matched characters that is, _ plus the following character. It won't take these two characters into consideration. Now [^_\\n]* matches any character but not of an _ or a \\n newline character zero or more times. So the characters after the character which was preceded by an underscore would be matched upto the next _ or \\n character. Removing those characters will give you the desired output.

DEMO

$input = preg_replace('~(?<=_.)[^_\n]*~', '', $input);

It just looks after to the _ and the character following the _ and matches all the characters upto the next underscore or newline character.

DEMO