Why is Python pandas limited to numpy.datetime64[ns]?

Question

pandas use of numpy.datetime64[ns] (datetime measured in nanoseconds, 10^-9 s) limits date (yyyy-mm-dd) from 1677-09-21 to 2262-04-11 and pushes to use either Python datetime library for AD dates or numpy.datetime[D] or Astropy for astronomical datetimes and julian days. Are stock trades performed on [ns] basis?

# numpy.datetime[D] (unit='D') is not recognized
import pandas as pd

pd.to_datetime(arg=(-753, 4, 21), yearfirst=True, unit='D')
# returns
DatetimeIndex(['1967-12-10', '1970-01-05', '1970-01-22'], dtype='datetime64[ns]', freq=None)

Answer 1

等待大神答复