Pandas Timestamp and .isin functionality

Question

I am trying to create a function that I will .apply() to a dataframe to:

• Remove one business day from the provided argument into the function
• Check to see if this new day is in a specific collection of dates (formatted as a datetime index)

I have simplified the logic of my function to iron out this issue - I will add more logic later.

My function:

def test(trade_date):
    if (trade_date - BDay(1)).isin(pricing_date):
        return True
    else:
        return False

Error:

AttributeError: 'Timestamp' object has no attribute 'isin'

Looks like there is an issue with using .isin with a Timestamp. However, when I run code to test within the dataframe itself:

df['Check'] = df['test_date'].isin(pricing_date)

The expected output is returned - isin() does work properly with this data.

TradeDate
2023-01-03    False
2023-01-03    False
2023-01-03    False
2023-01-03    False
2023-01-03    False
              ...  
2023-03-22     True
2023-03-22     True
2023-03-22     True
2023-03-22     True
2023-03-22     True
Name: Check, Length: 14324, dtype: bool

The column that .isin() is being called on is of datatype: datetime64[ns], but unsure how to convert the timestamp in my function to this data type - I have read in many places that they are virtually equivalent, just types from python vs pandas.

Name: test_date, Length: 14324, dtype: datetime64[ns]

Any help is appreciated!

Tried passing in a timestamp into .isin - expected output from running it directly on the dataframe.

Answer 1

Pandas dataframe apply runs the function on all the values inside the pd.Series rather than the function in the pd.Series. Thus, trade_date would be a timestamp, that doesn't have the isin method. What you should do is something like this:

def test(trade_date):
    return (trade_date - BDay(1)) in pricing_date

Or, much simpler:

df['Check'] = (df['test_date']-BDay(1)).isin(pricing_date)