Python Pandas - Return a new Timedelta with minutely floored resolution

To return a new Timedelta floored to this resolution, use the timedelta.floor() method. For minutely floored resolution, set the freq parameter to 'T'.

Syntax

timedelta.floor(freq)

Parameters

freq: String representing the frequency. Use 'T' or 'min' for minutes.

Creating a Timedelta Object

First, import pandas and create a Timedelta object ?

import pandas as pd

# Create a Timedelta object
timedelta = pd.Timedelta('8 days 11 hours 39 min 18 s')
print("Original Timedelta:")
print(timedelta)

Original Timedelta:
8 days 11:39:18

Applying Minutely Floor Resolution

Use the floor() method with freq='T' to floor to the nearest minute ?

import pandas as pd

# Create a Timedelta object
timedelta = pd.Timedelta('8 days 11 hours 39 min 18 s')
print("Original Timedelta:")
print(timedelta)

# Apply minutely floor resolution
floored_timedelta = timedelta.floor(freq='T')
print("\nTimedelta (minutely floored):")
print(floored_timedelta)

Original Timedelta:
8 days 11:39:18

Timedelta (minutely floored):
8 days 11:39:00

Multiple Examples with Different Frequencies

Compare different floor frequencies to understand the behavior ?

import pandas as pd

# Create a Timedelta with seconds
td = pd.Timedelta('2 days 5 hours 23 min 45 s')
print("Original:", td)

# Different floor frequencies
print("Minutely floor (T):", td.floor('T'))
print("Hourly floor (H):", td.floor('H'))
print("Daily floor (D):", td.floor('D'))

Original: 2 days 05:23:45
Minutely floor (T): 2 days 05:23:00
Hourly floor (H): 2 days 05:00:00
Daily floor (D): 2 days 00:00:00

Key Points

• The floor() method rounds down to the specified frequency
• Use 'T' or 'min' for minute-level flooring
• Seconds and microseconds are removed when flooring to minutes
• The original Timedelta object remains unchanged

Conclusion

The floor() method with freq='T' efficiently floors Timedelta objects to minute precision. This is useful for time-based data analysis where you need consistent minute-level granularity.