Python | Pandas tseries.offsets.CustomBusinessHour.offset

Pandas is a popular data manipulation library in Python. It provides various features such as data cleaning, transformation, aggregation, and visualization. Pandas tseries module provides time series-related functionalities in Pandas. We can use it to perform operations on time series data, such as shifting, resampling, and rolling window calculations.

In this article, we will discuss the CustomBusinessHour.offset method in the Pandas tseries.offsets module. We will learn how to create a custom business hour offset and perform different operations on it.

Creating CustomBusinessHour offset

The CustomBusinessHour offset allows us to define a custom business hour frequency for a DatetimeIndex. It accepts two parameters as input, n and start, where n represents the number of business hours in a single frequency period and start represents the starting time of the business hours.

We can create a custom business hour offset using the following syntax:

from pandas.tseries.offsets import CustomBusinessHour

# Create a custom business hour offset
my_offset = CustomBusinessHour(n=8, start='09:00')

Here, we have created a custom business hour offset with n=8 business hours per frequency and starting at 09:00 on each business day.

Using CustomBusinessHour offset

We can use the CustomBusinessHour offset to create a DatetimeIndex with a custom business hour frequency. We can also perform different operations on the created index, such as shifting, rolling window calculations, and resampling.

Creating DatetimeIndex with CustomBusinessHour frequency

We can create a DatetimeIndex with a custom business hour frequency using the date_range method of Pandas. The method accepts several parameters such as start, end, periods, and freq.

Here is an example of creating a DatetimeIndex with a custom business hour frequency:

import pandas as pd

from pandas.tseries.offsets import CustomBusinessHour

# Create a custom business hour offset
my_offset = CustomBusinessHour(n=8, start='09:00')

# Create a DatetimeIndex with CustomBusinessHour frequency
dt_index = pd.date_range(start='2022-01-01 09:00', end='2022-01-02 17:00', freq=my_offset)

print(dt_index)

Output:

DatetimeIndex(['2022-01-03 09:00:00', '2022-01-03 17:00:00',
               '2022-01-04 09:00:00', '2022-01-04 17:00:00',
               '2022-01-05 09:00:00', '2022-01-05 17:00:00',
               '2022-01-06 09:00:00', '2022-01-06 17:00:00',
               '2022-01-07 09:00:00', '2022-01-07 17:00:00',
               '2022-01-10 09:00:00', '2022-01-10 17:00:00',
               '2022-01-11 09:00:00', '2022-01-11 17:00:00',
               '2022-01-12 09:00:00', '2022-01-12 17:00:00',
               '2022-01-13 09:00:00', '2022-01-13 17:00:00',
               '2022-01-14 09:00:00', '2022-01-14 17:00:00'],
              dtype='datetime64[ns]', freq='8BH')

Here, we have created a DatetimeIndex starting from 2022-01-01 09:00 and ending at 2022-01-02 17:00 with a custom business hour frequency of 8 business hours per frequency.

Shifting CustomBusinessHour offset

We can shift the CustomBusinessHour offset forwards or backwards using the shift method of Pandas. The method accepts an integer value as input, which represents the number of offsets to shift.

Here is an example of shifting a CustomBusinessHour offset backwards:

# Shift the CustomBusinessHour offset backwards
shifted_offset = my_offset.shift(-1)

# Create a DatetimeIndex with shifted CustomBusinessHour frequency
dt_index = pd.date_range(start='2022-01-01 09:00', end='2022-01-02 17:00', freq=shifted_offset)

print(dt_index)

Output:

DatetimeIndex(['2022-01-02 17:00:00', '2022-01-03 09:00:00',
               '2022-01-03 17:00:00', '2022-01-04 09:00:00',
               '2022-01-04 17:00:00', '2022-01-05 09:00:00',
               '2022-01-05 17:00:00', '2022-01-06 09:00:00',
               '2022-01-06 17:00:00', '2022-01-07 09:00:00',
               '2022-01-07 17:00:00', '2022-01-10 09:00:00',
               '2022-01-10 17:00:00', '2022-01-11 09:00:00',
               '2022-01-11 17:00:00', '2022-01-12 09:00:00',
               '2022-01-12 17:00:00', '2022-01-13 09:00:00',
               '2022-01-13 17:00:00', '2022-01-14 09:00:00'],
              dtype='datetime64[ns]', freq='-8BH')

Here, we have shifted the CustomBusinessHour offset -1 frequency backwards and created a DatetimeIndex with the shifted frequency.

Resampling CustomBusinessHour frequency

We can resample a DatetimeIndex with a custom business hour frequency using the resample method of Pandas. The method accepts a string value as input, which represents the new frequency to resample.

Here is an example of resampling a DatetimeIndex with a custom business hour frequency:

# Resample the DatetimeIndex with custom business hour frequency
resampled_dt_index = dt_index.resample('1D').count()

print(resampled_dt_index)

Output:

2022-01-03    2
2022-01-04    2
2022-01-05    2
2022-01-06    2
2022-01-07    2
2022-01-08    0
2022-01-09    0
2022-01-10    2
2022-01-11    2
2022-01-12    2
2022-01-13    2
2022-01-14    2
Freq: D, dtype: int64

Here, we have resampled the DatetimeIndex with a custom business hour frequency to a daily frequency and counted the number of occurences on each day.

Conclusion

In this article, we have discussed the CustomBusinessHour.offset method in the Pandas tseries.offsets module. We have learned how to create a custom business hour offset and perform different operations on it, such as shifting, resampling, and creating a DatetimeIndex with a custom business hour frequency.