Python Pandas: How to Extract Hours, Minutes, and Seconds from Datetime in Pandas

Isolating time components from datetime columns is a common operation in data analysis. Whether you are grouping sales data by hour of the day, filtering log entries to business hours, or creating time-based features for a machine learning model, you need to extract individual time components from full datetime values. Pandas provides the .dt accessor for this purpose, offering fast vectorized access to hours, minutes, seconds, and other components without writing loops.

In this guide, you will learn how to extract time components from datetime columns, calculate total seconds, and apply these techniques to practical scenarios like hourly aggregation and time-based filtering.

Extracting Time Components with the .dt Accessor

The .dt accessor provides direct access to individual datetime components on an entire column at once. First, ensure your column is a proper datetime64 type, then access the components you need:

import pandas as pd

df = pd.DataFrame({
    'timestamp': ['2024-01-15 12:30:45', '2024-01-15 14:15:00', '2024-01-15 09:45:30']
})

# Convert strings to datetime
df['timestamp'] = pd.to_datetime(df['timestamp'])

# Extract individual components
df['hour'] = df['timestamp'].dt.hour
df['minute'] = df['timestamp'].dt.minute
df['second'] = df['timestamp'].dt.second

print(df)

Output:

            timestamp  hour  minute  second
0 2024-01-15 12:30:45    12      30      45
1 2024-01-15 14:15:00    14      15       0
2 2024-01-15 09:45:30     9      45      30

Each .dt attribute returns a Series of integers, making the results ready for arithmetic, grouping, or filtering.

Working with Time-Only Strings

When your data contains time values without dates, pd.to_datetime() still parses them correctly. Pandas assigns a default date, but the time components are extracted the same way:

import pandas as pd

df = pd.DataFrame({
    'time': ['12:30:45', '14:15:00', '09:45:30']
})

df['time'] = pd.to_datetime(df['time'])

df['hour'] = df['time'].dt.hour
df['minute'] = df['time'].dt.minute
df['second'] = df['time'].dt.second

print(df[['hour', 'minute', 'second']])

Output:

   hour  minute  second
0    12      30      45
1    14      15       0
2     9      45      30

Calculating Total Seconds Since Midnight

Converting a time to a single numeric value is useful for plotting, distance calculations, or creating features for analysis:

import pandas as pd

df = pd.DataFrame({
    'timestamp': pd.to_datetime(['12:30:45', '14:15:00', '00:05:30'])
})

df['total_seconds'] = (
    df['timestamp'].dt.hour * 3600 +
    df['timestamp'].dt.minute * 60 +
    df['timestamp'].dt.second
)

print(df[['timestamp', 'total_seconds']])

Output:

            timestamp  total_seconds
0 2026-02-17 12:30:45          45045
1 2026-02-17 14:15:00          51300
2 2026-02-17 00:05:30            330

12:30:45 translates to 45,045 seconds after midnight (12 times 3600 + 30 times 60 + 45).

Using total_seconds() with Timedelta Columns

For duration data stored as timedelta, the .dt.total_seconds() method provides a direct conversion:

import pandas as pd

df = pd.DataFrame({
    'duration': pd.to_timedelta(['1:30:45', '2:15:00', '0:45:30'])
})

df['seconds'] = df['duration'].dt.total_seconds()

print(df)

Output:

         duration  seconds
0 0 days 01:30:45   5445.0
1 0 days 02:15:00   8100.0
2 0 days 00:45:30   2730.0

info

The total_seconds() method is available only on timedelta columns, not on datetime columns. For datetime columns, calculate total seconds from midnight using the component-based formula shown above.

Available Time Component Attributes

Attribute	Description	Range
.dt.hour	Hour in 24-hour format	0 to 23
.dt.minute	Minute	0 to 59
.dt.second	Second	0 to 59
.dt.microsecond	Microsecond	0 to 999999
.dt.time	Time as Python time object	N/A

Practical Examples

Grouping Sales by Hour

Extracting the hour component lets you aggregate data by time of day to identify peak periods:

import pandas as pd

df = pd.DataFrame({
    'timestamp': pd.to_datetime([
        '2024-01-15 09:30:00', '2024-01-15 09:45:00',
        '2024-01-15 10:15:00', '2024-01-15 10:30:00'
    ]),
    'sales': [100, 150, 200, 175]
})

hourly = df.groupby(df['timestamp'].dt.hour)['sales'].sum()

print(hourly)

Output:

timestamp
9     250
10    375
Name: sales, dtype: int64

The 10 AM hour generated more sales (375) than the 9 AM hour (250).

Filtering by Time of Day

Use the extracted hour to filter records to specific time windows:

import pandas as pd

df = pd.DataFrame({
    'timestamp': pd.to_datetime([
        '2024-01-15 08:30:00', '2024-01-15 12:15:00',
        '2024-01-15 18:45:00', '2024-01-15 22:00:00'
    ]),
    'event': ['Login', 'Meeting', 'Logout', 'Alert']
})

# Morning events (before noon)
morning = df[df['timestamp'].dt.hour < 12]
print("Morning events:")
print(morning[['event', 'timestamp']])
print()

# Business hours (9 AM to 5 PM)
business = df[df['timestamp'].dt.hour.between(9, 17)]
print("Business hours events:")
print(business[['event', 'timestamp']])

Output:

Morning events:
   event           timestamp
0  Login 2024-01-15 08:30:00

Business hours events:
     event           timestamp
1  Meeting 2024-01-15 12:15:00

Creating a Time-of-Day Category

import pandas as pd
import numpy as np

df = pd.DataFrame({
    'timestamp': pd.to_datetime([
        '2024-01-15 06:00:00', '2024-01-15 10:00:00',
        '2024-01-15 15:00:00', '2024-01-15 21:00:00'
    ]),
    'activity': ['Jog', 'Work', 'Meeting', 'Dinner']
})

hour = df['timestamp'].dt.hour

df['period'] = np.select(
    [hour < 12, hour < 17, hour >= 17],
    ['Morning', 'Afternoon', 'Evening'],
    default='Unknown'  # <- string default
)

print(df[['activity', 'period']])

Output:

  activity     period
0      Jog    Morning
1     Work    Morning
2  Meeting  Afternoon
3   Dinner    Evening

tip

The .dt accessor works only on columns with datetime64 or timedelta64 dtypes. If your column is stored as strings, convert it first with pd.to_datetime() before accessing .dt attributes.

Quick Reference

Goal	Method
Extract hour	df['col'].dt.hour
Extract minute	df['col'].dt.minute
Extract second	df['col'].dt.second
Total seconds (timedelta)	df['col'].dt.total_seconds()
Time component only	df['col'].dt.time
Group by hour	df.groupby(df['col'].dt.hour)
Filter by time range	df[df['col'].dt.hour.between(9, 17)]

• Use the .dt accessor for fast, vectorized extraction of time components from datetime columns.
• Attributes like .dt.hour, .dt.minute, and .dt.second return integer Series that are immediately useful for grouping, filtering, and feature engineering.
• For duration data, use .dt.total_seconds() on timedelta columns to get a single numeric representation.