使用Pandas将嵌套的JSON转换为CSV，然后将CSV转换回JSON

Question

我有一个JSON主体，需要转换为CSV。然后，我需要从CSV进行完美的转换，以输出相同的JSON。因此，首先是JSON到CSV，然后是CSV到JSON

以下是JSON：

{
    "quoteId": "",
    "carrier": "",
    "productType": "",
    "effectiveDate": "",
    "referenceId": "",
    "agencyLocationCode": null,
    "locationId": null,
    "property": {
        "id": "",
        "address": {
            "propertyStreetNumber": "",
            "propertyStreetName": "",
            "propertyAddressLine2": null,
            "propertyCity": "",
            "propertyState": "",
            "propertyZipCode": ""
        },
        "details": null
    }
}

我运行以下代码得到的输出是：

输出csv:

代码：

with open('resources/json/file.json') as data_file:    
data = json.load(data_file)
df = pd.json_normalize(data)
df.to_csv("file.csv", index=False, encoding="utf-8")

现在，我希望将CSV转换回完全相同的JSON，目前我运行以下代码：

df = pd.read_csv('file.csv')
df.to_json('new.json')

我得到的输出JSON如下所示：

{
    "quoteId": {
        "0": null
    },
    "carrier": {
        "0": ""
    },
    "productType": {
        "0": ""
    },
    "effectiveDate": {
        "0": ""
    },
    "referenceId": {
        "0": 62646354
    },
    "agencyLocationCode": {
        "0": null
    },
    "locationId": {
        "0": null
    },
    "losses": {
        "0": null
    },
    "property.id": {
        "0": null
    },
    "property.address.propertyStreetNumber": {
        "0": ""
    },
    "property.address.propertyStreetName": {
        "0": ""
    },
    "property.address.propertyAddressLine2": {
        "0": null
    },
    "property.address.propertyCity": {
        "0": ""
    },
    "property.address.propertyState": {
        "0": ""
    },
    "property.address.propertyZipCode": {
        "0": ""
    },
    "property.details": ""
}

我需要输出json看起来像第一个。请帮忙