如何将“INVALID”分配给 function 的所有输出？

Question

I have to define a function that takes inputs Year, Crash_Month, Crash_Day and Crash_Time and returns the outputs season(Summer for months 12, 1, 2, Autumn for months 3,4,5 etc) weekday (Monday:1 to Sunday:7) and time_of_day (as Morning, Afternoon, Evening and Night). I have created the following code to define my function:

df = pd.DataFrame({'year': (data_dict['Year']),
            'season': (data_dict['Crash_Month']),
            'weekday': (data_dict['Crash_Day']),
            'time_of_day': (data_dict['Crash_Time'])})
df['time_of_day'] = pd.to_datetime(df['time_of_day'], format = "%H:%M:%S", errors = 'coerce')
d = {1:'Monday', 2:'Tuesday', 3:'Wednesday', 4:'Thursday', 5:'Friday', 6:'Saturday', 7:'Sunday'}
df['weekday'] = df['weekday'].map(d)
s = {12:'Summer', 1:'Summer', 2:'Summer', 3:'Autumn', 4:'Autumn', 5:'Autumn', 6:'Winter', 7:'Winter', 8: 'Winter', 9: 'Spring', 10: 'Spring', 11:'Spring'}
df['season'] = df['season'].map(s)
df['time_of_day']=df['time_of_day'].dt.hour.apply(lambda x: np.select([0<=x<6,
                                                                       6<=x<12,
                                                                       12<=x<18,
                                                                       12<=x<24],
                                                                      ['Night', 'Morning', 'Afternoon', 'Evening']))    
def compute_time_day_year(df):
    return(df)

It returns a data frame with the required outputs. However, in the excel file some of the Crash_Time values are "UNKNOWN". For these "UNKNOWN" entries I want to assign all outputs in the data frame (season, weekday, time) as 'INVALID". I have tried to use:

while True:
    try:
        compute_time_day_year(df)
    except ValueError:
        return 'INVALID' 

but the output does not change any values to INVALID. Does anyone have any advice on how to assign all output to 'INVALID'?

Answer 1

Because to_datetime has parameter errors = 'coerce' it return missing values for UNKNOWN strings, so then set NaN s to weekday and season columns:

df['time_of_day'] = pd.to_datetime(df['time_of_day'], format = "%H:%M:%S", errors = 'coerce')
df.loc[df['time_of_day'].isna(), ['weekday','season']] = np.nan

If want set all columns:

df.loc[df['time_of_day'].isna()] = np.nan
#or
#df = df.mask(df['time_of_day'].isna())