Python max() function fails when used on pandas columns and integer fails

Question

I'm trying to create a new pandas dataframe column by subtracting an existing dataframe column column from another. However, if the result is a negative number, the new column value should be set to zero.

import pandas as pd
data = {'A': [1,2,3], 'B': [3,2,1]}
df = pd.DataFrame(data)

In [4]: df
Out[4]: 
   A  B
0  1  3
1  2  2
2  3  1

If I create a new dataframe column 'C' by subtracting 'B' from 'A', I get the right result.

df['C'] = df['A'] - df['B']

In[8]: df
Out[7]: 
   A  B  C
0  1  3 -2
1  2  2  0
2  3  1  2

However, if I utilize the max() function to avoid results with a negative number, I get "ValueError: The truth value of a Series is ambiguous."

>>> df['C'] = max(df['A'] - df['B'], 0)
ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

The expected output is:

What am I doing wrong?

what is recommended output?

jezrael
– jezrael

2015-12-30 20:32:02 +00:00
Commented Dec 30, 2015 at 20:32 — jezrael
– jezrael, Commented Dec 30, 2015 at 20:32
In[8]: df Out[7]: A B C 0 1 3 0 1 2 2 0 2 3 1 2

vlmercado
– vlmercado

2015-12-30 20:33:48 +00:00
Commented Dec 30, 2015 at 20:33 — vlmercado
– vlmercado, Commented Dec 30, 2015 at 20:33

Alex Riley · Accepted Answer · 2015-12-30 20:33:03Z

13

You need to use np.maximum to do element-wise maximum comparison:

>>> np.maximum(df['A'] - df['B'], 0)
0    0
1    0
2    2
dtype: int64

The problem is max is that it essentially checks (df['A'] - df['B']) > 0. This returns an array of boolean values (not a boolean), hence the error.

answered Dec 30, 2015 at 20:33

Alex Riley

178k46 gold badges274 silver badges247 bronze badges

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

Comments

EdChum · Accepted Answer · 2015-12-30 20:33:14Z

1

Use np.where:

In [8]:
df['C'] = np.where((df['A'] - df['B'] > 0), df['A'] - df['B'], 0)
df

Out[8]:
   A  B  C
0  1  3  0
1  2  2  0
2  3  1  2

the in-built max function operates on scalars not array-like structures hence the error

answered Dec 30, 2015 at 20:33

EdChum

397k204 gold badges836 silver badges583 bronze badges

1 Comment

vlmercado Over a year ago

I used your recommended np.where suggestion somewhere else. Thanks.

efajardo · Accepted Answer · 2015-12-30 20:48:48Z

0

You could also do apply:

df['C'] = df.apply(lambda row: max(row['A'] - row['B'], 0), axis=1)

answered Dec 30, 2015 at 20:48

efajardo

8075 silver badges9 bronze badges

Comments

Mahyar Mottaghi Zadeh · Accepted Answer · 2021-01-31 17:20:34Z

0

Old post, but instead of using the max function:

df.max()

try to apply max-function on the values helped:

df.values.max()

edited Jan 31, 2021 at 17:20

Mahyar Mottaghi Zadeh

1,3256 gold badges21 silver badges33 bronze badges

answered Jan 31, 2021 at 15:32

user15117522

1

Collectives™ on Stack Overflow

Python max() function fails when used on pandas columns and integer fails

4 Answers 4

Comments

1 Comment

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

Comments

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related