Pandas：嵌套 json/dict 到 json_normalize DataFrame 和 json_normalize DataFrame 到嵌套 json/dict

Question

I have a large amount of JSON data and I want to perform some tasks. so I choose pandas for this.

I have a nested json like this:

json_data = [
    {
        "item": "Item1",
        "lowestPrice": {
            "price": 11.00,
            "currency": "EUR",
        },
    },
    {
        "item": "Item2",
        "lowestPrice": {
            "price": 12.00,
            "currency": "EUR",
        }
    },
    {
        "item": "Item3",
        "lowestPrice": {
            "price": 13.00,
            "currency": "EUR",
        }
    }
]

and i used json_normalize() to normalize nested json like:

df = pd.json_normalize(json_data, max_level=2)

        item  lowestPrice.price lowestPrice.currency
0  Item1               11.0                  EUR
1  Item2               12.0                  EUR
2  Item3               13.0                  EUR

#do something

now I need data back as a nested JSON or dict like:

json_data = [
    {
        "item": "Item1",
        "lowestPrice": {
            "price": 11.00,
            "currency": "EUR",
        },
        "annotatePrice": 15.00
    },
    {
        "item": "Item2",
        "lowestPrice": {
            "price": 12.00,
            "currency": "EUR",
        },
        "annotatePrice": 15.00
    },
    {
        "item": "Item3",
        "lowestPrice": {
            "price": 13.00,
            "currency": "EUR",
        },
        "annotatePrice": 15.00
    }
]

Answer 1

First, I added the column annotatePrice to the dataframe. Then constructed the inner dictionary for lowestPrice , followed by the outer dictionary. I sourced my solution from this stack answer .

Below is the dataframe after adding annotatePrice column.

Conversion code:

df = pd.json_normalize(json_data, max_level=2)
df['annotatePrice'] = 15

json_data = (df.groupby(['item', 'annotatePrice'])
       .apply(lambda x: x[['lowestPrice.price', 'lowestPrice.currency']].rename(columns={"lowestPrice.price":'price', "lowestPrice.currency":'currency'}).to_dict('records')[0])
       .reset_index()
       .rename(columns={0:'lowestPrice'})
       .to_dict(orient='records'))

json_data

Output:

[
  {
        'annotatePrice': 15,
        'item': 'Item1',
        'lowestPrice': {
            'currency': 'EUR',
            'price': 11.0
        }
    },
    {
        'annotatePrice': 15,
        'item': 'Item2',
        'lowestPrice': {
            'currency': 'EUR',
            'price': 12.0
        }
    },
    {
        'annotatePrice': 15,
        'item': 'Item3',
        'lowestPrice': {
            'currency': 'EUR',
            'price': 13.0
        }
    }
]