As titled.
I have written my code but it does not work. I wish I can get a more pythonic way of writing the code (in one single line perhaps).
clean_df :
columnA
123F
FVGD
w999Z
678Q
6y6yA
My code :
postfix = ["A", "D", "Z", "P"]
for value in postfix:
if cleaned_data['columnA'].str.endswith(value) is True:
cleaned_data['columnA'] = value
else:
cleaned_data['columnA'] = "blah"
The postfix are constant. Expected outcome :
columnA
blah
D
Z
blah
A
In one line with a list comprehension:
postfix = ["A", "D", "Z", "P"]
cleaned_data['columnA'] = [value[-1] if value[-1] in postfix else "blah" for value in cleaned_data['columnA']]
The output is :
columnA
blah
D
Z
blah
A
A simple one-liner would be
df['columnA'] = np.where(df.columnA.str[-1].isin(postfix), df.columnA.str[-1], 'blah')
np.where takes in a condition, value if the condition is True and value if the condition is False.
OR
In pure pandas, without using numpy, it would be
df['columnA'] = df.columnA.str[-1].where(df.columnA.str[-1].isin(postfix), 'blah')
You can try this with pd.Series.str.extract here.
pat = "|".join(postfix)
pat = f"({pat}$"
df['columnA'] = df['columnA'].str.extract(pat, expand=False).fillna('blah')
df
columnA
0 blah
1 D
2 Z
3 blah
4 A
You can use numpy.where:
Note: str.endswith accepts a tuple also.
In [3933]: import numpy as np
In [3934]: df.columnA = np.where(df.columnA.str.endswith(tuple(postfix)), df.columnA.str[-1], 'blah')
In [3935]: df
Out[3935]:
columnA
0 blah
1 D
2 Z
3 blah
4 A