time64 type is used to represent a specific moment in time. This
datetime in Python, or
timestamp in Arrow or pandas.
Internally, this type is stored as a 64-bit integer containing the number of
nanoseconds since the epoch (Jan 1, 1970) in UTC.
This type is not leap-seconds aware, meaning that it assumes that each day
has exactly 24×3600 seconds. In practice it means that calculating time
difference between two
time64 moments may be off by the number of leap
seconds that have occurred between them.
time64 type is not timezone-aware, addition of time zones is
planned for the next release.
A time64 column converts into
datetime.datetime objects in python, a
pa.timestamp('ns') type in pyarrow and
numpy and pandas.
DT = dt.Frame(["2018-01-31 03:16:57", "2021-06-15 15:44:23.951", None, "1965-11-25 19:29:00"]) DT = dt.Type.time64 DT