[Python] Since timezone info is deprecated by numpy.datetime64 constructor, what's the right...

I have a timeseries with a large number of ISO 8601 datestrings and would like to convert to numpy.datetime64. But I get a warning:

>>> import numpy as np
>>> t=np.datetime64('2022-05-01T00:00:00-07:00')
<stdin>:1: DeprecationWarning: parsing timezone aware datetimes is deprecated; this will raise an error in the future


OK, so how do I avoid the warning? (Without giving up a significant performance factor)

(Presumably the way to do this is convert to UTC and then create a datetime64 object, but the whole point of using datetime64 was that it worked directly on ISO 8601 strings very fast.)

The documentation mentions this issue but doesn't suggest any solution:

Deprecated since version 1.11.0: NumPy does not store timezone information. For backwards compatibility, datetime64 still parses timezone offsets, which it handles by converting to UTC±00:00 (Zulu time). This behaviour is deprecated and will raise an error in the future.
​

Continue reading...