CODEWITHSUNDEEP
javaregex
FunkSoulBrother
FunkSoulBrother

Reputation: 2167

Java : How to remove quotes from Json keys using regex

I need to strip the double quotes from a complex Json structure (object with children object and arrays in the hierarchy). I'm writing in Java (the app will run on Android).

I already have the string produced from a Json lib (Gson) and I am trying to strip the double quotes using regex instead of deserializing it back to an object and then serializing it without the double quotes in the key names.

I cannot find the right pattern or patterns to replace them.

example Json:

{
"key1":"value1",
"key2":"555f9ecd-2ced-42b7-b956-33e759bf8db1",
"key3":false,
"key4_parent":{
"child_key5":"Android",
"id":"7ae2facc-cdb3-7ae2-facccdb3bb86",
"name":"danny",
"bypassed":false
},
"object1":{
"deviceType":"Android",
"id":"7ae2facc-cdb3-7ae2-facccdb3bb86",
"base64val":"MENPSENJTFpZQzR3b1lvU1pQZTk=",
"name":"john"
},
"requiredLevel":"PUSH",
"level":"debug",
"status":1012,
"description":"Authentication in progress",
"objects":[
{
"deviceType":"Android",
"id":"7ae2facc-cdb3-7ae2-facccdb3bb86",
"base64val":"MENPSENJTFpZQzR3b1lvU1pQZTk=",
"name":"john"
},
{
"deviceType":"Android",
"id":"7ae2facc-cdb3-7ae2-facccdb3bb86",
"base64val":"MENPSENJTFpZQzR3b1lvU1pQZTk=",
"name":"john"
}
]
}

expected output :

{
key1:"value1",
key2:"555f9ecd-2ced-42b7-b956-33e759bf8db1",
key3:false,
key4_parent:{
child_key5:"Android",
id:"7ae2facc-cdb3-7ae2-facccdb3bb86",
name:"danny",
bypassed:false
},
object1:{
deviceType:"Android",
id:"7ae2facc-cdb3-7ae2-facccdb3bb86",
base64val:"MENPSENJTFpZQzR3b1lvU1pQZTk=",
name:"john"
},
requiredLevel:"PUSH",
level:"debug",
status:1012,
description:"Authentication in progress",
objects:[
{
deviceType:"Android",
id:"7ae2facc-cdb3-7ae2-facccdb3bb86",
base64val:"MENPSENJTFpZQzR3b1lvU1pQZTk=",
name:"john"
},
{
deviceType:"Android",
id:"7ae2facc-cdb3-7ae2-facccdb3bb86",
base64val:"MENPSENJTFpZQzR3b1lvU1pQZTk=",
name:"john"
}
]
}

I know that the output would not be a valid json. it's intended for easily searchable logging purposes.

any help would be greatly appreciated!

Upvotes: 9

Views: 8333

Answers (7)

Long Nguyễn Văn
Long Nguyễn Văn

Reputation: 69

Above, @toto's method is quite concise and precise. I can write it in detail this way:

text.replaceAll("\"([^\"]+)\":", "$1:")

Or you can try a more visual way like this:

String regex = "\"[a-z_]+\":";
final Pattern pattern = Pattern.compile(regex, Pattern.MULTILINE);
final Matcher matcher = pattern.matcher(text);

while (matcher.find()) {
    String keyField = matcher.group(0);
    String keyFieldWithoutQuotes = keyField.replace("\"", "");
    text = text.replaceFirst(keyField, keyFieldWithoutQuotes);
}
System.out.println(text);

This way it will find key fields containing quotes one by one, then delete the quotes.

Upvotes: 0

MorsBe
MorsBe

Reputation: 41

For those who come looking for a javascript answer and who need compatibility with all valid JSON including keys containing escaped quotes (\"), colons (:), or whitespace (\s), here are 2 options.

.replace(/("(\\[^]|[^\\"])*"(?!\s*:))|"((\\[^]|[^\\"])*)"(?=\s*:)/g, '$1$3')

OR

function unquoteKeys(json) {
  return json.replace(/"(\\[^]|[^\\"])*"\s*:?/g, function (match) {
    if (/:$/.test(match)) return match.replace(/^"|"(?=\s*:$)/g, '');
    else return match;
  }
}

This type of function can be used to do almost anything with valid JSON.

function modifyJson(json) {
  return json.replace(/"(\\[^]|[^\\"])*"(\s*:)?|\b(true|false|null)\b|-?\d+(\.\d+)?([eE][-+]?\d+)?/g, function (match) {
    if (/^"/.test(match)) {
      if (/:$/.test(match)) {
        // match = Key with enclosing quotes & following colon (".*"\s*:)
      } else {
        // match = String (not key) with enclosing quotes (".*")
      }
    } else if (/true|false/.test(match)) {
      // match = Boolean value (true|false)
    } else if (/null/.test(match)) {
      // match = Null value (null)
    } else {
      // match = Number (-?\d+(\.\d+)?([eE][-+]?\d+)?)
    }
  });
}

Upvotes: 4

Roma Khomyshyn
Roma Khomyshyn

Reputation: 1152

Could you try it?

jsonStr.replaceAll("\"(\\w+)\":", "$1:"));

Regex remove quotes from keys

Upvotes: 2

Vitalii Andrusishyn
Vitalii Andrusishyn

Reputation: 4162

try something like this:

jsonStr.replace(/\"/g, '');

Upvotes: 0

Tamas Rev
Tamas Rev

Reputation: 7166

Assuming there is no colon (:) in the jsonstr, except as key-value separator, you can experiment with lookaheads ((?=...) constructs) too:

jsonStr.replaceAll("\"(?=.*:)", "");

Upvotes: 0

Yu Jiaao
Yu Jiaao

Reputation: 4714

You use find by regex:

\"(\w+)\"\:

and replace with

$1:

In any editor

and in java will be 

String res = str.replaceAll("\\\"(\\w+)\\\"\\:","\\\\$1:")

Upvotes: 0

Toto
Toto

Reputation: 91385

I'd do:

  • find "([^"]+)":
  • replace $1:

Upvotes: 20

Related Questions