Pandas lambda multiple argument in return

Question

import pandas as pd
import numpy as np

def ced(x):
    return x+1, x+2, x+3


df = pd.DataFrame(data=[[1,2],[10,20]], columns=['a','b'])


df['x'], df['y'], df['z'] = df['a'].apply(lambda x: ced(x)) 


print(df)

error:

line 11, in df['x'], df['y'], df['z'] = df['a'].apply(lambda x: ced(x)) ValueError: not enough values to unpack (expected 3, got 2)

this thing works for

import pandas as pd
import numpy as np

def ced(x):
    return x+1, x+2


df = pd.DataFrame(data=[[1,2],[10,20]], columns=['a','b'])


df['x'], df['y'] = df['a'].apply(lambda x: ced(x)) 


print(df)

output:

    a   b  x   y
0   1   2  2  11
1  10  20  3  12

I don't know what is the problem here.

jezrael · Accepted Answer · 2018-07-11 10:42:36Z

1

I suggest change function for return Series and subset of new columns:

def ced(x):
    return pd.Series([x+1, x+2, x+2])

df = pd.DataFrame(data=[[1,2],[10,20]], columns=['a','b'])


df[['x','y', 'z']] = df['a'].apply(lambda x: ced(x)) 

print(df)
    a   b   x   y   z
0   1   2   2   3   3
1  10  20  11  12  12

Another solution is create DataFrame by constructor:

def ced(x):
    return x+1, x+2, x+2

df = pd.DataFrame(data=[[1,2],[10,20]], columns=['a','b'])


df[['x','y', 'z']] = pd.DataFrame(df['a'].apply(lambda x: ced(x)).values.tolist())
print(df)
    a   b   x   y   z
0   1   2   2   3   3
1  10  20  11  12  12

answered Jul 11, 2018 at 10:42

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Nihal Over a year ago

can you tell which one will be faster first one or the second for large dataframe?

jezrael Over a year ago

@Nihal - I think second one.

jezrael Over a year ago

@Nihal - Just tested

In [260]: %timeit df[['x','y', 'z']] = pd.DataFrame(df['a'].apply(lambda x: ced(x)).values.tolist()) 3.05 ms ± 99.1 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) In [262]: %timeit df[['x','y', 'z']] = df['a'].apply(lambda x: ced(x)) 349 ms ± 8.56 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

jezrael Over a year ago

with df = pd.concat([df] * 1000, ignore_index=True)

Collectives™ on Stack Overflow

Pandas lambda multiple argument in return

1 Answer 1

4 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related