Change Pandas JSON output structure

Question

I have a Pandas DataFrame, that after applying a groupby function

df.groupby(['USER', 'GROUP'])['VALOR'].sum()

outputs this:

USER               GROUP
John Doe           A        201.37
                   B        480.59
                   C       1504.16
John Jones         A        239.95
                   B       1123.39
                   C       1736.05
...

However, when I convert this df to JSON:

df.to_json()

It returns this:

{"["John Doe","A"]":201.37,"["John Doe","B"]":480.59,"["John Doe","C"]":1504.16", "["John Jones","A"]":239.95,"["John Jones","B"]":1123.39,"["John Jones","C"]":1736.05}

Is there a way to output it in this format:

{"John Doe": {"A": 201.37, "B":480.59, "C":1504.16}, "John Jones": {"A": 239.95, "B":1123.39, "C":1736.05}}

Answer 1

Yes, unstack it first.

df.unstack(0).to_json()

{
    "John Doe": {
        "A": 201.37,
        "B": 480.59,
        "C": 1504.16
    },
    "John Jones": {
        "A": 239.95,
        "B": 1123.39,
        "C": 1736.05
    }
}