Python Pandas - Return index locations of values at particular time of day in DateTimeIndex

To return index locations of values at particular time of day in DateTimeIndex, use the DateTimeIndex.indexer_at_time() method. This method returns an array of integer positions where the time component matches the specified time.

Syntax

DateTimeIndex.indexer_at_time(time, asof=False)

Parameters

The key parameters are ?

• time ? Time as a time object or string
• asof ? Return the latest index location if exact time not found (default: False)

Creating DateTimeIndex

First, let's create a DateTimeIndex with timezone-aware timestamps ?

import pandas as pd

# Create DatetimeIndex with 20-minute intervals
datetimeindex = pd.date_range('2021-10-30 02:30:50', periods=5, tz='Australia/Adelaide', freq='20T')

print("DateTimeIndex...")
print(datetimeindex)
print("\nFrequency:", datetimeindex.freq)

DateTimeIndex...
DatetimeIndex(['2021-10-30 02:30:50+10:30', '2021-10-30 02:50:50+10:30',
               '2021-10-30 03:10:50+10:30', '2021-10-30 03:30:50+10:30',
               '2021-10-30 03:50:50+10:30'],
              dtype='datetime64[ns, Australia/Adelaide]', freq='20T')

Frequency: <20 * Minutes>

Finding Index Locations by Time

Use indexer_at_time() to get the integer positions of timestamps matching a specific time ?

import pandas as pd

datetimeindex = pd.date_range('2021-10-30 02:30:50', periods=5, tz='Australia/Adelaide', freq='20T')

# Find index locations for specific time
locations = datetimeindex.indexer_at_time('2021-10-30 03:10:50')
print("Index locations at 03:10:50:", locations)

# You can also use just the time part
locations_time_only = datetimeindex.indexer_at_time('03:10:50')
print("Using time only:", locations_time_only)

Index locations at 03:10:50: [2]
Using time only: [2]

Multiple Matches Example

When multiple timestamps have the same time component, all matching positions are returned ?

import pandas as pd

# Create index spanning multiple days with same times
dates = ['2021-10-30 15:30:00', '2021-10-31 15:30:00', '2021-11-01 15:30:00', '2021-11-02 10:00:00']
multi_day_index = pd.DatetimeIndex(dates)

print("Multi-day DateTimeIndex:")
print(multi_day_index)

# Find all positions with time 15:30:00
locations = multi_day_index.indexer_at_time('15:30:00')
print("\nPositions with time 15:30:00:", locations)

Multi-day DateTimeIndex:
DatetimeIndex(['2021-10-30 15:30:00', '2021-10-31 15:30:00',
               '2021-11-01 15:30:00', '2021-11-02 10:00:00'],
              dtype='datetime64[ns]', freq=None)

Positions with time 15:30:00: [0 1 2]

Comparison with Related Methods

Conclusion

The indexer_at_time() method efficiently returns integer positions of timestamps matching a specific time component. Use this method when you need index positions rather than the actual datetime values for further operations.