CODEWITHSUNDEEP
pandasdatetimetime
cphill
cphill

Reputation: 5914

Average datetime.time Series

I am trying to figure out the best way to average a series of datetime.time values with about 40 records in the format of 23:19:30 or HH:MM:SS, however, when I attempt to use the to_timedelta method and apply .mean() I run into an error:

ValueError: Invalid type for timedelta scalar: <class 'datetime.time'>

What would be the proper format for the time to be in to use to_timedelta().mean() or is there a better approach?

Code:

bed_time_mean = pd.to_timedelta(bed_rise_df['start_time']).mean()

Upvotes: 0

Views: 163

Answers (1)

FObersteiner
FObersteiner

Reputation: 25544

pd.to_timedelta accepts strings in H:M:S format as input, so you could convert your column with datetime.time objects to string first. Ex:

from datetime import time
import pandas as pd

df = pd.DataFrame({'t':[time(1,2,3), time(2,3,4), time(3,4,5)]})

pd.to_timedelta(df['t'].astype(str)).mean()
# Timedelta('0 days 02:03:04')

Upvotes: 1

Related Questions