AttributeError: 'TimedeltaProperties' object has no attribute 'minute'

Question

I have a dataframe that looks like this

df

[output]:
date        time
2020-02-28  00:30:45
2020-02-28  00:30:45
2020-03-09  00:21:06
2020-03-09  00:21:06
2020-03-09  00:21:06

with

df.time.dtype

[output]: dtype('<m8[ns]')

I want to extract the minutes in the time variable with the following command

df.time.dt.minute

but instead, I have this error

AttributeError: 'TimedeltaProperties' object has no attribute 'minute'

Does someone know how to fix this problem?

Answer 1

your column 'time' is of dtype timedelta as the error tells you; you could use the total_seconds() method to convert to seconds and divide by 60 to get the minutes.

If you want a full-featured datetime column, combine 'date' and 'time'. Then you can use .dt.minute.

Ex:

import pandas as pd
df = pd.DataFrame({'time': pd.to_timedelta(['00:30:45','00:30:45','00:21:06','00:21:06','00:21:06']),
                   'date': pd.to_datetime(['2020-02-28','2020-02-28','2020-03-09','2020-03-09','2020-03-09'])})

# to get the "total minutes":
df['minutes'] = df['time'].dt.total_seconds()/60
df['minutes']
# 0    30.75
# 1    30.75
# 2    21.10
# 3    21.10
# 4    21.10
# Name: minutes, dtype: float64

[pd.Timedelta docs]

# to get a column of dtype datetime:
df['DateTime'] = df['date'] + df['time']

# now you can do:
df['DateTime'].dt.minute
# 0    30
# 1    30
# 2    21
# 3    21
# 4    21
# Name: DateTime, dtype: int64

Answer 2

If you have not converted to a datetime dataframe do that first then you create a new column like this

df['minute'] = df['date'].dt.minute

or this method here

  df[new]= df[column].map(lambda x: datetime.datetime(x.minutes))

Answer 3

@Fobersteiner's answer is very good, but just for completeness, I would like to add that you could also divide your column of dtype timedelta by a fixed timedelta. For instance:

from datetime import timedelta
import pandas as pd
df = pd.DataFrame({'time': pd.to_timedelta(['00:30:45','00:30:45','00:21:06','00:21:06','00:21:06']),
                   'date': pd.to_datetime(['2020-02-28','2020-02-28','2020-03-09','2020-03-09','2020-03-09'])})

# to get the "total minutes":
df['minutes'] = df['time'] / timedelta(minutes=1)  # <--
df['minutes']

Out[9]: 
0    30.75
1    30.75
2    21.10
3    21.10
4    21.10
Name: minutes, dtype: float64

Though personally, I prefer @Fobersteiner's method.