熊猫聚合后删除组头

Question

Suppose I have the following:

In [20]: df.groupby("dummy").agg({"returns": [np.mean, np.sum]})
Out[20]: 
        returns          
            sum      mean

dummy                    
1      0.285833  0.028583

How do I remove the "returns" header? So that I have a dataframe with columns 'sum' and 'mean'?

Answer 1

You don't even to remove the level after grouping if you can avoid creating it in the first place. Call agg on your column of interest -

df.groupby("dummy")['returns'].agg(['mean', 'sum'])

---

Demo

df = pd.DataFrame({'dummy':[1, 1, 2, 2], 'returns' : [50, 50, 100, 200]})
df
   dummy  returns
0      1       50
1      1       50
2      2      100
3      2      200

df.groupby("dummy")['returns'].agg(['mean', 'sum'])

       mean  sum
dummy           
1        50  100
2       150  300

Alternatively, using numpy's functions -

df.groupby("dummy")['returns'].agg([np.mean, np.sum])

       mean  sum
dummy           
1        50  100
2       150  300

Answer 2

Use -

df.columns = df.columns.droplevel(0)

To drop the top column level

Answer 3

result = df.groupby("dummy").agg({"returns": [np.mean, np.sum]})
result.columns = result.columns.get_level_values(1)