Open Pabloendres opened 3 years ago
I'm having the same issue. Perhaps you knew this already, but it seems that this check:
if len(np.unique(np.diff(s.index))) > 1:
ends up counting days even though the series has a monthly frequency. df.index.is_monotonic
allows for the frequency so it sees an increment of 1 for each month, whereas the numpy approach results in an array of 28, 29, 30, and 31 days so the len() function is 4, which is > 1.
Hi! I'm working with monthly data from 2006 to 2019, and I wanted to work with SeasonalAD, but it fails with "ValueError: The time steps are not constant." even after validating the series.
Code
Output
Checks