Transform first level JSON Key with Python - Solved

user_9976350 0
Viewed 122

I have the following JSON:

{
    "1605855600000":
    [
        {
            "id": "1",
            "value": "1. Choice 1",
            "checked": true
        },
        {
            "id": "2",
            "value": "2. Choice 2",
            "checked": false
        },
        {
            "id": "3",
            "value": "3. Choice 3",
            "checked": false
        },
        {
            "id": "4",
            "value": "4. Choice 4",
            "checked": false
        },
        {
            "id": "5",
            "value": "5. Choice 5",
            "checked": false
        }
    ],
    "1604732400000":
    [
        {
            "id": "1",
            "value": "1. Choice 1",
            "checked": false
        },
        {
            "id": "2",
            "value": "2. Choice 2",
            "checked": false
        },
        {
            "id": "3",
            "value": "3. Choice 3",
            "checked": false
        },
        {
            "id": "4",
            "value": "4. Choice 4",
            "checked": false
        },
        {
            "id": "5",
            "value": "5. Choice 5",
            "checked": false
        }
    ]
}

I'm not sure the exact terminology but the "keys" like '1605855600000' and '1604732400000' are UNIX timestamps that I need to convert in Python.

Here is what I have so far:

def convert_timestamp(obj):
    for key in obj.keys():
        timestamp = int(key) / 1000
        dt_object = datetime.fromtimestamp(timestamp)
        date = dt_object.strftime("%B %d, %Y")
        new_key = date
        if new_key != key:
            obj[new_key] = obj[key]
            del obj[key]
    return obj

data = json.loads(data, object_hook=convert_timestamp)

However, the error I receive says: ValueError: invalid literal for int() with base 10: 'id' which means that it is accessing a level below what I am trying to change.

I know the logic inside the definition works to output the string that I need as the final result, but accessing and replacing those with the current logic isn't working.

The final output needs to look something like:

{
    "12 November, 2020":
    [
        {
            "id": "1",
            "value": "1. Choice 1",
            "checked": true
        },
...

Here's how I eventually solved it.

json_data = json.loads(data)
dict_keys = json_data.keys()

timestamp = []
for x in dict_keys:
    date_val = int(x) / 1000
    dt_object = datetime.fromtimestamp(date_val)
    date = dt_object.strftime("%B %d, %Y")
    timestamp.append(date)

final_dict = dict(zip(timestamp, list(json_data.values())))

Essentially it ended up being easier to pull the keys into a list, perform the transformation, and then merge that list back to the values as the new keys instead of trying to do it in one fell swoop. For someone who has strict performance requirements this might not be ideal but for this use case it will work.

python-3.x python json
3 Answers

You are much better off making a new dict than messing with the keys:

import json
from datetime import datetime

data_str = '''{
    "1605855600000":
    [
        { "id": "1", "value": "1. Choice 1", "checked": true },
        { "id": "2", "value": "2. Choice 2", "checked": false },
        { "id": "3", "value": "3. Choice 3", "checked": false },
        { "id": "4", "value": "4. Choice 4", "checked": false },
        { "id": "5", "value": "5. Choice 5", "checked": false }
    ],
    "1604732400000":
    [
        { "id": "1", "value": "1. Choice 1", "checked": false },
        { "id": "2", "value": "2. Choice 2", "checked": false },
        { "id": "3", "value": "3. Choice 3", "checked": false },
        { "id": "4", "value": "4. Choice 4", "checked": false },
        { "id": "5", "value": "5. Choice 5", "checked": false }
    ]
}'''

def convert_timestamp(key):
    timestamp = int(key) / 1000
    dt_object = datetime.fromtimestamp(timestamp)
    return dt_object.strftime("%B %d, %Y")

def convert_keys(obj):
    return {convert_timestamp(key):val for (key,val) in obj.items()}

data = json.loads(data_str)

new_data = convert_keys(data)
print(new_data)

Output:

{'November 20, 2020': [{'id': '1', 'value': '1. Choice 1', 'checked': True}, {'id': '2', 'value': '2. Choice 2', 'checked': False}, {'id': '3', 'value': '3. Choice 3', 'checked': False}, {'id': '4', 'value': '4. Choice 4', 'checked': False}, {'id': '5', 'value': '5. Choice 5', 'checked': False}], 'November 07, 2020': [{'id': '1', 'value': '1. Choice 1', 'checked': False}, {'id': '2', 'value': '2. Choice 2', 'checked': False}, {'id': '3', 'value': '3. Choice 3', 'checked': False}, {'id': '4', 'value': '4. Choice 4', 'checked': False}, {'id': '5', 'value': '5. Choice 5', 'checked': False}]}
quamrana_48340

Did you mean:

data = json.loads(data)

data = convert_timestamp(data)
quamrana_48340

You misused object_hook parameter in json.loads.

From the docs

object_hook is an optional function that will be called with the result of any object literal decoded (a dict). The return value of object_hook will be used instead of the dict. This feature can be used to implement custom decoders (e.g. JSON-RPC class hinting).

Therefore, it takes the first object found by the key. As @quamrana mentioned in his answer you need to do:

json_data = json.loads(data)
converted_data = convert_timestamp(json_data)

Also, it is considered a good practice not to override data but use separate variables.

user_44733860
Related