schmidt-ai opened a new issue, #45531: URL: https://github.com/apache/arrow/issues/45531
### Describe the bug, including details regarding any error messages, version, and platform. Goal: create a fixed shape tensor column with `dim_names` from a batch of data. Within `FixedShapeTensorArray.from_numpy_ndarray` [here](https://github.com/apache/arrow/blob/main/python/pyarrow/array.pxi#L4537), `fixed_shape_tensor` is called. I think `from_numpy_ndarray` should take a `dim_names` argument (or `**kwargs`) and pass it to `from_numpy_ndarray`. Repro: ```python import numpy as np import pyarrow as pa tensor_type = pa.fixed_shape_tensor(value_type=pa.float32(), shape=(5, 6), dim_names=("a", "b")) schema = pa.schema([("tensor", tensor_type)]) batch = {"tensor": np.random.rand(10, 5, 6)} batch["tensor"] = pa.FixedShapeTensorArray.from_numpy_ndarray(batch["tensor"]) pa.Table.from_pydict(batch, schema=schema) ``` Error: `pyarrow.lib.ArrowTypeError: Casting from 'extension<arrow.fixed_shape_tensor[value_type=float, shape=[5,6], permutation=[0,1]]>' to different extension type 'extension<arrow.fixed_shape_tensor[value_type=float, shape=[5,6], dim_names=[a,b]]>' not permitted. One can first cast to the storage type, then to the extension type.` > One can first cast to the storage type, then to the extension type. It's not clear how exactly to do this. I tried adding `batch["tensor"] = batch["tensor"].cast(tensor_type)` to the snippet above but got the same result. Perhaps I'm missing something? ### Component(s) Python -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@arrow.apache.org.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
