pandas groupby: an overall/total row?

问题

I have the following:

maths = {'floatval': 'mean'}
out = df.groupby(['pid', 'year', 'month']).agg(maths)

Which gives me a grouped mean of floatval per year and month, which is great, but I want an all/overall row, which has the mean val of floatval across a row just grouped by pid. Should I make a second data frame and combine, or is there a more efficient way?

来源：https://stackoverflow.com/questions/51712739/pandas-groupby-an-overall-total-row