正则表达式将字符串“to”中的所有双引号替换为有效的json字符串

Question

I have a invalid json string like {"subject":"tv 60" new","body":"contact me for detail"}.

And i try to use php preg_replace to valid this json string but fail.

Please help me to create a regex to replace the " to /" .

Thanks in advance.

Answer 1

Assuming that the '{', ':' and the ',' characters in your JSON don't have any trailing space between them and the double quotes, a solution could be this:

$sanitizedJsonString = preg_replace('/([^:,{])"([^:,}])/', "$1".'\"'."$2", $yourMalformedJsonString);

It finds all the blockquotes not preceded by the '{', ':' or ',' characters and not followed by the ':', ',' or '}' characters and replaces them with the '\\"' character.


Hopefully it solve your problems, however keep in mind that there are some special cases in which this approach fails: for example, if inside your malformed JSON you have a word inside blockquotes that is followed by a comma like this:

{"subject":"tv 60" new","body":"contact "me", for detail"}

you'll still have a not valid JSON:

{"subject":"tv 60\" new","body":"contact \"me", for detail"}

If you need a more reliable solution I agree with Steve P., you have to implement a script that parses your string and make the replacements of the blockquotes when required.

Answer 2

$string =~ /(?!<[\{:,])\"(?!\s*[:,\}]/\\"/g  

这个正则表达式查找任何“不遵循a：a或a {并且不跟随a：a或a}并且用\\字符引用它。这应该适合你。

Answer 3

Unless you're regular expressions to validate the keys/values prior to insertion into the JSON object, using regular expressions is not a good way to attempt to validate a given JSON object.

Even if you are attempting to validate upon insertion, it seems like regex is overkill and instead you can just use various helper methods or simply iterate through the raw string and replace when necessary.