熊猫groupby函数的总和不超过8列

Question

I have a pandas dataframe that contains 13 text columns and 16 numeric columns (29 columns in total, around 13k rows). I would like to aggregate the data by the first 13 columns and return the sum of the results for the 16 numeric columns. I have tried the following:

df.groupby(1,2,3,4,5,6,7,8,9,10,11,12,13)[14,15,26,17,18,19,20,21,22,23,24,25,26,27,28,29].sum()

but this returns an error "groupby() takes from 1 to 8 positional arguments but 14 were given"

I am essentially trying to do the following as it would be in SQL syntax:

select 1,2,3,4,5,6,7,8,9,10,11,12,13,sum(14),sum(15),sum(16),sum(17),sum(18),sum(19),sum(20),sum(21),sum(22),sum(23),sum(24),sum(25),sum(26),sum(27),sum(28),sum(29)
from df group by 1,2,3,4,5,6,7,8,9,10,11,12,13

I'd also like the process done in place so I end up with a dataframe the same shape as the old one (with fewer rows, obviously!)

any help appreciated, thanks!

Answer 1

A slightly more general approach that uses .select_dtypes ( docs ) to isolate numeric columns:

import pandas as pd
import numpy as np

numerical_columns = df.select_dtypes(include=[np.number]).columns.tolist()
other_columns = df.select_dtypes(exclude=[np.number]).columns.tolist()

df.groupby(other_columns)[numerical_columns].sum()

As for why your code is not working, it is because you need to pass a list of column names to groupby.