如何将多索引数据帧（按多列分组的数据帧）转换为嵌套的 json

Question

My Pandas Series, which I got from applying groupby operation on DataFrame with columns 'var' and 'month' and applying sum on the corresponding data looks like this ('var' and 'month' are indexes below) :

    var  month
X   Feb     -0.061575
    Jan      1.366478
Y   Feb     -1.310896
Z   Apr      0.053076
    Feb      1.292415
    Mar      0.375144
P   Feb      1.241288
    Mar      0.613453

What I want a format of JSON created from the above DataFrame like below:

'data':[{'label': 'X', 'data': ['Jan': 1.366478, 'Feb': -0.061575]}, ... ]

I know the basic pandas .to_json() may not work here. Probably a combination of list comprehension, lambda function etc. can work here?

The closest I could think of is :

dict = {k: df[k].to_dict() for k in df.index.levels[0]}

This produce {'X': {'Feb': -0.06157474257929787, 'Jan': 1.366478487212244},'Y': ...}

Any help is appreciated.

Thanks

Answer 1

'data':[{'label': 'X', 'data': ['Jan': 1.366478, 'Feb': -0.061575]}, ... ]

This is an invalid json. The inner list doesn't make sense

Answer 2

For me the solution I found is below piece of code (Assuming group_data holds the already grouped by data on a DataFrame).

group_dict = {k: group_data[k].to_dict() for k in group_data.index.levels[0]}
group_list = []
for k, v in group_dict.items():
    dict = {'label': k, 'data': v}
    group_list.append(dict)