IndexedOptionArray.to_IndexedOptionArray64 fails when index dtype is int32 #3726
Description
Version of Awkward Array
2.8.10
Description and code to reproduce
Calling IndexedOptionArray.to_IndexedOptionArray64() on an array backed by an Index32 raises AttributeError. The implementation passes the Index wrapper directly to the backend astype, but only the underlying NumPy array supports that method. Index.to64() already performs the correct conversion and can be used internally to fix this.
import numpy as np
import awkward as ak
from awkward.index import Index
index32 = Index(np.array([0, 1, -1], dtype=np.int32))
content = ak.contents.NumpyArray(np.array([10, 20, 30]))
arr = ak.contents.IndexedOptionArray(index32, content)
arr.to_IndexedOptionArray64()Traceback (most recent call last):
File "/data/src/test.py", line 10, in <module>
arr.to_IndexedOptionArray64()
File "/home/hdd/miniconda3/envs/py312/lib/python3.12/site-packages/awkward/contents/indexedoptionarray.py", line 272, in to_IndexedOptionArray64
self._backend.nplike.astype(self._index, dtype=np.int64),
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/hdd/miniconda3/envs/py312/lib/python3.12/site-packages/awkward/_nplikes/array_module.py", line 751, in astype
return x.astype(dtype, copy=copy) # type: ignore[attr-defined]
^^^^^^^^
AttributeError: 'Index32' object has no attribute 'astype'. Did you mean: 'dtype'?