delete \ in column of obejct when covert to json pandas python

Question

i have dataframe User

id | user  | phone
------------------
1  | chris | {"device": "a", "brand": "a"}
2  |  nat  | {"device": "a", "brand": "a"}
3  | zack  | {"device": "a", "brand": "a"}

then pass it to object variable

rows = df.loc[df['id'] == 1]
merged = {'data': rows.reindex(columns=df.columns.values).to_dict('records')}
with open((path + '/' + filename + '.json'), 'w') as fp:
        json.dump(merged, fp)

then got filename.json:

{
  "data": {
    "id" : 1,
    "user" : "chris",
    "phone": "{\"device\": \"a\", \"brand\": \"a\"}"
   }
}

expected result:

{
  "data": {
    "id" : 1,
    "user" : "chris",
    "phone": {
        "device": "a", 
        "brand": "a"
     }
   }
}

how to make it as expected result?

Answer 1

The content of phone is already JSON. Converting JSON to JSON will lead to escaping the quotes, as you are experiencing.

The solution is to turn the JSON string of phone to a dict first using json.loads ), before you convert merged to JSON.

import json

merged = {
  "data": {
    "id" : 1,
    "user" : "chris",
    "phone": "{\"device\": \"a\", \"brand\": \"a\"}"
   }
}

merged['data']['phone'] = json.loads(merged['data']['phone'])

print(json.dumps(merged, indent=4))

Output:

{
    "data": {
        "id": 1,
        "user": "chris",
        "phone": {
            "device": "a",
            "brand": "a"
        }
    }
}

Of course, this assumes that phone is always valid JSON, otherwise json.loads will fail.