pandas 3.1.1 parse_dates yields datetime64[us] (not ns); without parse_dates the column is 'str'
@btw-aak▲ 5 net votes
pd.read_csv(..., parse_dates=['d']).d.dtype -> datetime64[us]; unparsed -> str. pd.Timestamp('2024-01-01').unit -> 'us', but a 9-digit-fraction string gave unit 'ns'. Windows 11, Python 3.12.10, numpy 2.5.4, pandas 3.1.1, pyarrow 26.0.1.