Getting ValueError when using pandas lambda with multiple inputs and outputs

Question

When executing the code below

import numpy as np
import pandas as pd

def somef(t,y,u):
# Some silly function but close enough to represent actual use case
    y=t+u+y
    g=1*u*t
    l,o=y[0],g[0]
    return l,o
data=np.array([4,6,1,3])

df=pd.DataFrame(dict(a=[1,2,3,4],b=[3,6,4,1]))
df[['ttt','uuu']]=df.apply(lambda x: somef(data,x['a'],x['b']), axis=1)

The compiler return an error

ValueError: Columns must be same length as key

May I know what is the issue?

Expected output

a   b   ttt uuu
1   3   8   12
2   6   12  24
3   4   11  16
4   1   9   4

Answer 1

Pandas apply have an argument result_type . Assigning the result_type equal to expand resolve the issue.

'expand': list-like results will be turned into columns.

Such that,

df[['ttt','uuu']]=df.apply(lambda x: somef(data,x['a'],x['b']), axis=1,result_type="expand")

Return

   a  b  ttt  uuu
0  1  3    8   12
1  2  6   12   24
2  3  4   11   16
3  4  1    9    4

Interesting enough, setting the output column as df['ttt','uuu'] and the result_type as None (which is the default value )

df['ttt','uuu']=df.apply(lambda x: somef(data,x['a'],x['b']), axis=1)

Return

   a  b (ttt, uuu)
0  1  3    (8, 12)
1  2  6   (12, 24)
2  3  4   (11, 16)
3  4  1     (9, 4)