Format Problem Categorizing Time In Pandas

I'm trying to convert hours to an categoric format, the column is like this with hundreds of entries. hr_animals 21:25:00 10:36:00 23:17:00 01:23:00 NA 13:30:00 NA And i want the

Solution 1:

That's probably because your data is not datetime type so you cannot use .dt access. To fix your code, the 3rd option should be:

pd.to_datetime(pamdf['hr_animals'], format='%H:%M:%S', errors='coerce').dt.hour

That said, your format is better suited for timedelta instead of DateTime. Try:

df['hr_animals'] = ((pd.to_timedelta(df['hr_animals'], errors='coerce')
                      // pd.Timedelta('4H') )
                      .add(1)
                      .replace({1: 'Dawn',
                      2: 'Early Morning',
                      3: 'Morning',
                      4: 'Noon',
                      5: 'Evening',
                      6: 'Night'})
                   )

Output:

  hr_animals
0      Night
1    Morning
2      Night
3       Dawn
4NaN5       Noon
6NaN

Another option is to use pd.cut, which returns a categorical column. This might be helpful because the labels will be ordered, i.e. Dawn < Early Morning<...:

df['hr_animals'] = pd.cut(pd.to_timedelta(df['hr_animals'], errors='coerce'),
                          bins = pd.to_timedelta(np.arange(0,25,4), unit='H'),
                          labels=['Dawn','Early Morning', 'Morning', 
                                  'Noon', 'Evening', 'Night']
                   )