dt64d.astype(“<M8[us]”) dt64ns.astype(“<M8[D]”)
Remarks:
NumPy can store date and times in multiple formats inside datetime64.
The stored int64, which can be accessed using value property may be different even if the represented date is the same.
Related: