PHP Regex replace all instances of a character in similar strings - php

Question

I have a file that contains a collection strings. All of the strings begin with the same set of characters and end with the same character. I need to find all of the strings that match a certain pattern, and then remove particular characters from them before saving the file. Each string looks like this:
Data_*: " ... "
where Data_ is the same for each string, the asterisk is an incrementing integer that is either two or three digits, and the colon and the double quotation marks are the same for each string. The ... is completely different in every string and it's the part of each I need to work with. I need to remove all double quotation marks from the ... , preserving the enclosing double quotation marks. I don't need to replace them, just remove them.
So for example, I need this...
Data_83: "He said, "Yes!" to the question"
to become this...
Data_83: "He said, Yes! to the question"
I am familiar with PHP and would like to use this. I know how to do something like...
<?php
$filename = 'path/to/file';
$content = file_get_contents($filename);
$new_content = str_replace('"', '', $content);
file_put_contents($filename, $new_content);
And I'm pretty sure a regular expression will be what I'm wanting to use to find the strings and remove the extra double quotation marks. But I'm very new to regular expressions and need some help here.
EDIT:
I should have mentioned, the file is a PHP file containing an object. It looks a bit like this:
<?php
$thing = {
Data_83: "He said, "Yes!" to the question",
Data_84: "Another string with "unwanted" quotes"
}

Answer 1

You may use preg_replace_callback with a regex like
'~^(\h*Data_\d{2,}:\h*")(.*)"~m'
Note that you may make it safer if you specify an optional , at the end of the line: '~^(\h*Data_\d{2,}:\h*")(.*)",?\h*$~m' but you might need to introduce another capturing group then (around ,?\h*, and then append $m[3] in the preg_replace_callback callback function).
Details
^ - start of the line (m is a multiline modifier)
(\h*Data_\d{2,}:\h*") - Group 1 ($m[1]):
\h* - 0+ horizontal whitespaces
Data_ - Data_ substring
\d{2,} - 2 or more digits
: - a colon
\h* - 0+ horizontal whitespaces
" - double quote
(.*) - Group 2 ($m[2]): any 0+ chars other than line break chars, as many as possible, up to the last...
" - double quote (on a line).
The $m represents the whole match object, and you only need to remove the " inside $m[2], the second capture.
See the PHP demo:
preg_replace_callback('~^(\h*Data_\d{2,}:\h*")(.*)"~m', function($m) {
return $m[1] . str_replace('"', '', $m[2]) . '"';
}, $content);

Answer 2

Not as elegant but you could create a UDF:
function RemoveNestedQuotes($string)
{
$firstPart = explode(":", $string)[0];
preg_match('/"(.*)"/', $string, $matches, PREG_OFFSET_CAPTURE);
$tmpString = $matches[1][0];
return $firstPart . ': "' . preg_replace('/"/', '', $tmpString) . '"';
}
example:
$string = 'Data_83: "He said, "Yes!" to the question"';
echo RemoveNestedQuotes($string);
// Data_83: "He said, Yes! to the question"

Answer 3

One more step after str_replace with implode and explode. You can just do it like this.
<?php
$string = 'Data_83: "He said, "Yes!" to the question"';
$string = str_replace('"', '', $string);
echo $string =implode(': "',explode(': ',$string)).'"';
?>
Demo : https://eval.in/912466
Program Output
Data_83: "He said, Yes! to the question"
Just to replace " quotes
<?php
$string = 'Data_83: "He said, "Yes!" to the question"';
echo preg_replace('/"/', '', $string);
?>
Demo : https://eval.in/912457

Answer 4

The way I see it, you don't need to make any preg_replace_callback() calls or a convoluted run of explosions and replacements. You merely need to disqualify the 2 double quotes that you wish to retain and match the rest for removal.
Code: (Demo)
$string = 'Data_83: "He said, "Yes!" to the question",
Data_184: "He said, "WTF!" to the question"';
echo preg_replace('/^[^"]+"(*SKIP)(*FAIL)|"(?!,\R|$)/m','',$string);
Output:
Data_83: "He said, Yes! to the question",
Data_184: "He said, WTF! to the question"
Pattern Demo
/^[^"]+"(*SKIP)(*FAIL)|"(?!,?$)/m
This pattern says:
match from the start of each line until you reach the first double quote, then DISQUALIFY it.
then after the |, match all double quotes that are not optionally followed by a comma then the end of line.
While this pattern worked on regex101 with my sample input, when I transferred it to the php sandbox to whack together a demo, I needed to add \R to maintain accuracy. You can test to see which is appropriate for your server/environment.