python-将嵌套的json规范化为pandas数据帧

Question

I'm trying to normalize some JSON to flatten it for a SQL table. The problem I've come across is that everything I've read has a standard name for nested items but I am working with unique ids that I want to put in as values in my dataframe.

Here's a sample of the JSON .

{
"data": {
    "138082239": [
        {
            "id": 275,
            "name": "Sue",
            "abbreviation": "SJ",
            "active": true,
            "changedByUserId": "11710250",
            "statusUpdated": "2020-11-23T18:48:28+00:00",
            "leadCreated": "2020-11-23T18:48:28+00:00",
            "leadModified": "2020-11-23T18:48:29+00:00"
        }
    ],
    "138082238": [
        {
            "id": 276,
            "name": "John",
            "abbreviation": "JC",
            "active": true,
            "changedByUserId": "11710250",
            "statusUpdated": "2020-11-23T18:48:25+00:00",
            "leadCreated": "2020-11-23T18:48:25+00:00",
            "leadModified": "2020-11-23T18:48:25+00:00"
        }
    ],

I want to flatten this and add the index title (ex: 138082239) as a value [LeadId] in my dataframe. When I try to use pd.json_normalize() I just get a bunch of columns titled; data.138082239, data.138082238, etc.

I'm using requests to pull this JSON from an API.

    r = requests.request("GET", url, data=payload, headers=headers)
    
    j = r.json()
    
    df = pd.json_normalize(j)

I want the dataframe to look like this:

LeadId      id    name   abbreviation   active
138082239   275   Sue    SJ             TRUE
138082238   276   John   JC             TRUE

Thanks in advance for your help!

Answer 1

One way using dict comprehension with pandas.DataFrame.assign :

df = pd.concat([pd.DataFrame(v).assign(LeadId=k) for k, v in j["data"].items()])
print(df[["LeadId", "id", "name", "abbreviation", "active"]])

Output:

      LeadId   id  name abbreviation  active
0  138082239  275   Sue           SJ    True
0  138082238  276  John           JC    True