Call the `dt.` type from a Python Function

Question

I have timestamps in pandas and I want to be able to write a function that allows me to pull back different parts of the date. Here is an MWE:

import pandas as pd
df = pd.DataFrame({'hit': pd.to_datetime(['2018-12-29 00:00:00', '2019-03-25 00:00:00'])})

And this is what I want to be able to do:

def test(timepart):
    df['hit'].dt.timepart

So I could call test(month) and the function would return 12, 03.

Answer 1

Use getattr to return that particular attribute:

def test(df, timepart):
    return getattr(df['hit'].dt, timepart)

test(df, 'month')
0    12
1     3
Name: hit, dtype: int64

The issue with strftime is that it converts the result to a string before returning it.

---

If you want a taste of the dynamic life, try DataFrame.eval instead.

def test(df, timepart):
    return df.eval(f"hit.dt.{timepart}", engine='python')

test(df, 'month')

0    12
1     3
Name: hit, dtype: int64

python's eval is dangerous, thankfully there are safer alternatives to do the same thing.

Answer 2

You can use strftime:

def test(timepart):
    return df['hit'].dt.strftime('%'+timepart[0])
print(test('month'))

Output:

0    12
1    03
Name: hit, dtype: object