python – Combine structured JSON with remaining JSON

May 7, 2024

I have this structured df (json) which includes another json and I want to combine them to access all the values:

{
  "index": "exp-000005",
  "type": "_doc",
  "score": 9.502488,
  "source": {
    "verb": "REPLIED",
    "timestamp": "2022-01-20T08:14:00+00:00",
    "in_context": {
      "screen_width": "3440",
      "screen_height": "1440",
      "build_version": "7235",
      "question": "Hallo",
      "request_time": "403",
      "status": "success"
    }
  }
}

The result I want is the following:

{
  "index": "exp-000005",
  "type": "_doc",
  "score": 9.502488,
  "verb": "REPLIED",
  "timestamp": "2022-01-20T08:14:00+00:00",
  "screen_width": "3440",
  "screen_height": "1440",
  "build_version": "7235",
  "question": "Hallo",
  "request_time": "403",
  "status": "success"
}

How can I combine this structured json to one unstructured json? In R I used tbl_json. Maybe there is something similar for python?

>Solution :

Looks like you want to flatten the dict and only pick the values of the keys which are themselves dictionary.

Here is a recursive approach:

from pprint import pprint

d = {
    "index": "exp-000005",
    "type": "_doc",
    "score": 9.502488,
    "source": {
        "verb": "REPLIED",
        "timestamp": "2022-01-20T08:14:00+00:00",
        "in_context": {
            "screen_width": "3440",
            "screen_height": "1440",
            "build_version": "7235",
            "question": "Hallo",
            "request_time": "403",
            "status": "success",
        },
    },
}


def flatten(d: dict, result_d={}):
    for k, v in d.items():
        if isinstance(v, dict):
            flatten(v, result_d)
        else:
            result_d[k] = v
    return result_d


new_d = flatten(d)

pprint(new_d, sort_dicts=False)

output:

{'index': 'exp-000005',
 'type': '_doc',
 'score': 9.502488,
 'verb': 'REPLIED',
 'timestamp': '2022-01-20T08:14:00+00:00',
 'screen_width': '3440',
 'screen_height': '1440',
 'build_version': '7235',
 'question': 'Hallo',
 'request_time': '403',
 'status': 'success'}