Applying lambda on Python DataFrame

Question

I am having a Python Pandas DataFrame like

>>> df
  classification like 
0         flower    1 
1         flower    0 
2         flower    0 
3      adventure    1 
4      adventure    1

I want to create an output DataFrame like

>>> df
  classification like  liked
0         flower    1  True
1         flower    0  False
2         flower    0  False
3      adventure    1  True
4      adventure    1  True

I am "apply"ing the Python lambda function on the input DataFrame as follows:

>>> df['like'].apply(lambda x: x == 1)

But I am getting all 'False' under the 'liked' column

>>> df
  classification like  liked
0         flower    1  False
1         flower    0  False
2         flower    0  False
3      adventure    1  False
4      adventure    1  False

Any quick suggestions will be helpful.

>>> df['like'].astype(int)
0    1
1    0
2    0
3    1
4    1
Name: like, dtype: int32

@jezrael

>>> df['liked'] = df['like'].astype(bool)
>>> df
  classification like  liked
0         flower    1   True
1         flower    0   True
2         flower    0   True
3      adventure    1   True
4      adventure    1   True

@jezrael : DTypes

>>> df.dtypes
classification    object
like              object
liked               bool
dtype: object

If you can avoid `apply` which is a hidden loop as your needs can be vectorized. — Parfait, Jun 29 '18 at 13:14

jezrael · Accepted Answer · 2018-06-29T13:25:16.800

4

Convert integer column to boolean:

print (type(df.loc[0, 'like']))
<class 'numpy.int64'>

df['liked'] = df['like'].astype(bool)

Or assign comparison with 1:

df['liked'] = df['like'] == 1

If 1 is string compare by string '1':

print (type(df.loc[0, 'like']))
<class 'str'>

df['liked'] = df['like'] == '1'
print (df)
  classification like  liked
0         flower    1   True
1         flower    0  False
2         flower    0  False
3      adventure    1   True
4      adventure    1   True

edited Jun 29 '18 at 13:25

answered Jun 29 '18 at 13:13

jezrael

729,927
78
1,141
1,090

When I compared using df['like'] == '1' it is working. But the problem is when I checked the data type for the 'like' column it is showing as int32 – somnathchakrabarti Jun 29 '18 at 13:24
Do you check it by `print (type(df.loc[0, 'like']))` ? – jezrael Jun 29 '18 at 13:25
No I checked by df['like'].astype(int) – somnathchakrabarti Jun 29 '18 at 13:26
Here is possible check it by `print (df.dtypes)`, it return `object` for string column – jezrael Jun 29 '18 at 13:27

Applying lambda on Python DataFrame

1 Answers1