When working with pandas dataframes in Python, it is often necessary to add business days to a column of dates. However, this task becomes more complex when we need to skip over holidays. In this article, we will explore three different ways to solve this problem.

Option 1: Using the pandas CustomBusinessDay class

The first option involves using the pandas CustomBusinessDay class, which allows us to define our own business day calendar. We can specify the holidays to skip over and then use the offset method to add business days to the dates.

import pandas as pd
from pandas.tseries.offsets import CustomBusinessDay

# Define the holidays to skip over
holidays = ['2022-01-01', '2022-12-25']

# Create a CustomBusinessDay object with the specified holidays
bday = CustomBusinessDay(holidays=holidays)

# Add business days to the dates column
df['new_dates'] = df['dates'] + pd.offsets.BDay(n=5, holidays=bday)

This solution allows us to easily skip over holidays and add business days to the dates column. However, it requires importing the pandas CustomBusinessDay class and may not be the most efficient option for large datasets.

Option 2: Using the numpy busday_offset function

The second option involves using the numpy busday_offset function, which calculates the number of business days from a given date. We can specify the holidays to skip over and then use this function to add business days to the dates.

import pandas as pd
import numpy as np

# Define the holidays to skip over
holidays = ['2022-01-01', '2022-12-25']

# Add business days to the dates column
df['new_dates'] = df['dates'] + pd.to_timedelta(np.busday_offset(df['dates'].values.astype('datetime64[D]'), n=5, holidays=holidays), unit='D')

This solution leverages the numpy busday_offset function to calculate the number of business days and then adds the corresponding timedelta to the dates column. It is a more efficient option for large datasets but requires converting the dates column to numpy datetime64 format.

Option 3: Using a custom function with a loop

The third option involves creating a custom function that iterates over the dates and adds business days while skipping over holidays. This solution is more flexible and allows for custom logic, but it may be slower for large datasets.

import pandas as pd

# Define the holidays to skip over
holidays = ['2022-01-01', '2022-12-25']

# Custom function to add business days and skip over holidays
def add_business_days(date, n):
    for _ in range(n):
        date += pd.DateOffset(days=1)
        while date.weekday() in [5, 6] or date.strftime('%Y-%m-%d') in holidays:
            date += pd.DateOffset(days=1)
    return date

# Add business days to the dates column using the custom function
df['new_dates'] = df['dates'].apply(lambda x: add_business_days(x, 5))

This solution provides the most flexibility as we can customize the logic for skipping over holidays. However, it involves iterating over each date in the dataframe, which may be slower for large datasets.

After evaluating the three options, the best choice depends on the specific requirements of the task. If efficiency is a priority and the dataset is large, option 2 using the numpy busday_offset function is recommended. However, if flexibility and custom logic are more important, option 3 with a custom function may be the better choice. Option 1 using the pandas CustomBusinessDay class is a good balance between efficiency and flexibility.