bug: datetime64[ns] and datetime64[s]

[peterdudfield]

I found a different when using datetime64[ns] and datetime64[s]

pvlib version == 0.10.2
numpy version == 1.23.2
This was on a Mac Big Sur 11.3.1
python 3.10

import pvlib
import numpy as np

dt = np.array([datetime(2024,1,31, 12,0), datetime(2024,1,31,12,30)], dtype='datetime64[s]')

solpos = pvlib.solarposition.get_solarposition(
            time=dt,
            latitude=50,
            longitude=0)

Perhaps I don't fully understand the different between [ns] and [s] but would be great if anyone can help on this

openclimatefix/ocf_datapipes#272

[cwhanse]

Thanks @peterdudfield this results from pandas >= 2.0 #1932 and is being accomodated in pvlib with #1948

[peterdudfield]

Thanks @cwhanse for the quick response

[kandersolar]

Duplicate of #1932, fixed by #1948