How do i Flip a regex?

Question

i am using the following regex to check input for anyting other then the allowed characters...(a-zA-z0-9 a dot, comma, dash and a singlequote and the br tag)

<?php

$regex_char_appearance = '/([A-Za-z0-9 \-\.\,\']|(<br>))/';

?>

I have been trying to create a regex to clean user input. I just can't get it to work so tried different things like...

<?php

    $regex_char_appearance = '/(?!<br>)([^A-Za-z0-9 \-\.\,\'])/';

    $regex_char_appearance = '/([^A-Za-z0-9 \-\.\,\']|[^(<br>)])/';

   //remove anything other then alphabetic and allowed
    $post_char_appearance = preg_replace( $regex_char_appearance , '' , $post_char_appearance);

?>

so the goal is to remove anything other then a-zA-z0-9, a dot, comma, dash , singlequote and the br tag for output with preg_replace.

Can someone help me put a regex together that works?

I guess `[^A-Za-z0-9.,'-]` should be enough for the *remove anything other then a-zA-**Z**0-9, a dot, comma, dash and a singlequote* requirement. What is `
` doing here? You want to match anything other than `
`? — Wiktor Stribiżew, Apr 20 '16 at 08:19
Ok, then just match it and restore with a backreferences: `preg_replace('~(
)|[^A-Za-z0-9.,'-]~' , '$1' , $post_char_appearance);` - [demo](https://regex101.com/r/yJ0yU8/2) — Wiktor Stribiżew, Apr 20 '16 at 08:38
nope..it doesn't leave the
intact...thanks for the demo...usefull! — Joe Boss, Apr 20 '16 at 08:57
Please post your example code proving it does not work. As you see in my demo, the `
` are reinserted back into string while `preg_replace`ing. — Wiktor Stribiżew, Apr 20 '16 at 08:58
Actually...that worked! I never used a callback before but i will read into it...thanks! — Joe Boss, Apr 20 '16 at 09:14

score 0 · Accepted Answer · edited May 23 '17 at 12:16

You can match your disallowed characters with [^A-Za-z0-9.,'-] (a negated character class matching any character but those defined in the class).

To keep br tags intact, match and capture it with a pair of unescaped parentheses (a grouping construct (...)) and restore with a backreference $1:

$regex_char_appearance = '~(<br\s*/?>)|[^A-Za-z0-9.,'-]~';
$post_char_appearance = preg_replace($regex_char_appearance, '$1' , $post_char_appearance);

See the regex demo

Note that [A-z] does not only match all ASCII letters, see more on this in Why is this regex allowing a caret?

How do i Flip a regex?

1 Answers1