-
Notifications
You must be signed in to change notification settings - Fork 99
Open
Labels
bug (unverified)The problem described would be a bug, but needs to be triagedThe problem described would be a bug, but needs to be triaged
Description
Version of Awkward Array
2.8.3
Description and code to reproduce
Dear developers,
I found that for some reasons, when processing int arrays, ak.std returns nan, even though np.std does return a value. I have checked that there is no nan or inf in the array. Here are the codes to reproduce the problem:
import awkward as ak
import numpy as np
data = np.loadtxt("data.csv", delimiter=",", dtype=np.int64) # A 1D array
print("awkward:", ak.std(data))
print("numpy:", np.std(data))The output would be
awkward: nan
numpy: 458469074.0604589
Data file is as followed:
My understanding is that ak.std should perform as np.std, at least should be the same when being applied on 1D array, so I think this can be a bug.
The platform information is: Python 3.11 on AlmaLinuxOS9.
Hope this report can be helpful.
Best!
Metadata
Metadata
Assignees
Labels
bug (unverified)The problem described would be a bug, but needs to be triagedThe problem described would be a bug, but needs to be triaged