Remove an object from JSON using RegEx

Question

I have JSON objects in this format:

 {
     "1f626": {
         "name": "frowning face with open mouth",
         "ascii": [],
         "code_points": {
             "base": "1f626",
             "default_matches": [
                 "1f626"
             ],
             "greedy_matches": [
                 "1f626"
             ],
             "decimal": ""
         }
     }
 }

I have to remove the code_points object using Regular Expressions.

I have tried using this RegEx:

(("code\w+)(.*)(}))

But it is only selecting the first line. I have to select until end of curly brackets in order to fully get rid of the code_points object.

How can I do that?

Note: I have to remove it using Regular Expressions and not JavaScript. Please don't post any JavaScript answers or mark this as a possible duplicate of a JavaScript-based question.

Yes. Just try `delete obj["1f626"]["code_points"]`, with `obj` being the object in your code. — KaiserKatze, Aug 26 '18 at 04:09
Reference: [1](https://stackoverflow.com/questions/6485127/how-to-delete-unset-the-properties-of-a-javascript-object); [2](https://stackoverflow.com/questions/208105/how-do-i-remove-a-property-from-a-javascript-object); [3](https://stackoverflow.com/questions/1596782/how-to-unset-a-javascript-variable). — KaiserKatze, Aug 26 '18 at 04:12
If so, you are asking for [Lexical analysis](https://en.wikipedia.org/wiki/Lexical_analysis#Tokenization) feature. — KaiserKatze, Aug 26 '18 at 04:18
JSON isn't a regular language; it is actually awful to use regex on JSON and it is why we have JSON parsers. I dread to think who is forcing you to use regex :-( — Chris Cousins, Aug 26 '18 at 04:18
Possible duplicate of [How to implement Lexical Analysis in Javascript](https://stackoverflow.com/questions/4726539/how-to-implement-lexical-analysis-in-javascript). — KaiserKatze, Aug 26 '18 at 04:22
If you have to do it without javascript, maybe the javascript tag on the question isn't appropriate? — caedmon, Aug 26 '18 at 04:23
I hope you understand that **JSON encoder/decoder is not written in Regular Expression**. — KaiserKatze, Aug 26 '18 at 04:29
@Mina You could choose to use a [**JSON encoder/decoder written in C/C++**](https://json.org/). — KaiserKatze, Aug 26 '18 at 04:32
this is a large JSON file: https://raw.githubusercontent.com/delowar64/emoji-finder/master/src/emoji/emoji.json I have to reduce file size, that's why I want to use Regular expression. @KaiserKatze — Mina, Aug 26 '18 at 04:32
@Mina I don't understand why you choose not to use JavaScript. Efficiency? — KaiserKatze, Aug 26 '18 at 04:34
@KaiserKatze if I use javascript method then how can get output value as a new file content? — Mina, Aug 26 '18 at 04:40
@Mina If you use JavaScript to handle JSON, you should learn to use [**Node.js**](https://nodejs.org/en/), which is a standalone JavaScript Engine. — KaiserKatze, Aug 26 '18 at 04:43
@Mina You could try to use [**Python**](https://www.python.org) to handle JSON as well. — KaiserKatze, Aug 26 '18 at 04:43

Tom Blodget · Accepted Answer · 2018-08-27T22:51:33.547

3

Alternatively, at the command-line, if you can use jq

jq "del(.[].code_points)" <monster.json >smaller_monster.json

This deletes the code_points key inside each 2nd-level object.

It took my machine about 5 seconds on a 60MB document.

It's not a regular expression but it's not JavaScript, either. So, it meets half of your non-functional requirements.

edited Aug 27 '18 at 22:51

answered Aug 26 '18 at 23:12

Tom Blodget

20,260
3
39
72

thanks for your answer, this code also removing object key & comma: http://prntscr.com/knh22y here is code snippet: https://jqplay.org/s/ohqeX8OnG_ – Mina Aug 27 '18 at 16:16
can you please help me about this? – Mina Aug 27 '18 at 16:16
@Mina. Weird. I fixed the query. – Tom Blodget Aug 27 '18 at 22:52

score 1 · Answer 2 · answered Aug 26 '18 at 08:38

("code_points")([\s\S]*?)(})

The problem you had is that . is actually any character except \n, so in this case I usually use [\s\S] which means any whitespace and non-whitespace character (so it's actually any character). Also you should make * quantifier to be lazy by adding ?.

Remember that this Regular Expression won't work properly in case you have inner object (other {}) in code_points

Remove an object from JSON using RegEx

2 Answers2