Regular Expression to find \"

Question

What's the regular expression to find \"

I think it's this: '/\\"/' but I need to use it on a really large dataset so need to make sure this is correct.

I need to replace it with " so my code is : $data = preg_replace('/\\"/', '"', $data)

Is that correct?

Answer 1

For matching backslashes you need to 'double-escape' them, so you have four \ at the end:

$data = preg_replace('/\\\\"/', '"', $data);

Why you need 4 \: PHP parses a string \\" as \" and RegEx interprets this as " since in RegEx you don't need to escape ". So it wont match \". \\\\" will be parsed as \\" which will be interpreted as \" by RegEx.

Answer 2

A backslash does not need to be escaped in either a single-quoted string or a regular expression, unless the following character is a character that can be escaped (such as the backslash itself).

A double quote does not need to be escaped and cannot be escaped in a single-quoted string. In a regular expression it doesn't have to be either, but it can be.

That means \\ in both a single-quoted string and a regular expression becomes \, while \" in a single-quoted string remains \", while in a regular expression it becomes ".

However, in PHP you can only create a regular expression from a string, so you have to escape twice.

In other words...

Original string    String processed   Regexp processed
'/\"/'             /\"/               "
'/\\"/'            /\"/               "
'/\\\"/'           /\\"/              \"
'/\\\\"/'          /\\"/              \"
'/\\\\\"/'         /\\\"/             \"
'/\\\\\\"/'        /\\\"/             \"
'/\\\\\\\"/'       /\\\\"/            \\"

Bonus backslash

In a double-quoted string, of course, the " does need to be escaped, so...

"/\"/"             /"/                "
"/\\"/"            syntax error
"/\\\"/"           /\"/               "
"/\\\\"/"          syntax error
"/\\\\\"/"         /\\"/              \"
"/\\\\\\"/"        syntax error
"/\\\\\\\"/"       /\\\"/             \"
"/\\\\\\\\"/"      syntax error
"/\\\\\\\\\"/"     /\\\\"/            \\"

I think you should probably go for preg_replace("/\\\\\\\"/", "\"", $data) just to be on the safeconfusing side.