Converting dates when importing from CSV, OutOfBoundsDatetime: Out of bounds nanosecond timestamp. Pandas

Question

I'm importing data from a csv, and I'm trying to set a specific date to today's date.

Data in the csv if formatted this way:

All data in that column are dates and are formatted exactly the same. I read in the data with df = pd.read_csv(r'<filapath.csv>) at the moment.

Then this is run to convert all instances of '7/21/2020' into today's date:

df['filedate'] = np.where(pd.to_datetime(df['filedate']) == '7/21/2020', pd.Timestamp('now').floor(freq='d'),df['filedate'])

I receive this error: pandas._libs.tslibs.np_datetime.OutOfBoundsDatetime: Out of bounds nanosecond timestamp: 1-01-14 00:00:00

I don't want to use errors='coerce' because the column will always be 100% populated with real dates, and I will later need to filter the dataframe by date. There seems to be some "ghost" precision in the csv data I can't see. I cannot modify the csv column in this case and I can't use any packages outside of pandas and numpy.

Answer 1

...or alternatively .loc :

df.loc[df['filedate'] == '7/21/2020', 'filedate'] = pd.Timestamp('now').floor(freq='d')

Answer 2

Use.replace() function.

df['filedate'].replace({'7/21/2020':pd.Timestamp('now').floor(freq='d')})