[FIXED] How to create a calculated column function based on conditions on another column

Issue

I am trying to create new a calculated date column by subtracting two date columns, but only if a condition in a 3rd column (column_name) is met (it cannot be NaN). This is the code I have so far but it is not working (see error below). Thank you in advance as I am a beginner.

def function(column_name):
    if df[df.column_name.notna()] == True:
        return (df['date1']-df['date2']).astype('timedelta64[s]')
    else:
        return 'Null'

df['New_Calculated_Column'] = df['columnname'].apply(function)

Received this error:

Unexpected exception formatting exception. Falling back to standard exception

Solution

You can use pd.Series.where, e.g. with df on the 1st floor. Here's the code:

df['new']=df['col'].where(df['col'].isna(),df['date1'].sub(df['date2']).astype('timedelta64[s]'))

Answered By - cxywz