Open habr-ioki opened 8 months ago
Interesting... is the dtype of your input series int64? Could you try casting it to float64?
Amazing, that was the reason for this error. Thank you for this hint! Now it's working as expected.
May I suggest referring to this in the esda.G_Local docu? It only says that the y
parameter should be of type array. Perhaps you could specify this as an array of floats.
I think that we should check the dtype and cast it ourselves under the hood if it comes as int.
Indeed, we should do this casting to float64 ourselves for all statistics that hit the conditional inference engine
Hey, I'm getting following error when calling
esda.getisord.G_Local(df["column"], w)
.Input params are of type
As far as I understand the input types are correct and I also don't get any dimension mismatch errors.
Unfortunately I don't really understand the error message, but it seems to me that there's sth. not working as expected with numba (
numba.core.errors.TypingError: Failed in nopython mode pipeline (step: nopython frontend)
).Can anyone help me with this issue?
Python version: 3.11 libpysal: 4.9.2 pysal: 23.7 esda: 2.5.1