Could not read the data in MinIO by lance-ray

When I use the following code

import ray
import lance_namespace as ln
from lance_ray import read_lance, write_lance

# Initialize Ray
ray.init()

# Connect to a metadata catalog (directory-based example)
namespace = ln.connect("rest", {xxx})

# Create a Ray dataset
data = ray.data.range(1000).map(lambda row: {"id": row["id"], "value": row["id"] * 2})

# Write to Lance format using metadata catalog
write_lance(data, namespace=namespace, table_id=["lance_minio_catalog", "schema","my_table32"], storage_options = {'lance.storage.access_key_id' :  'x','lance.storage.endpoint' : 'http://minio:9000','lance.storage.secret_access_key' : 'x','lance.storage.allow_http' : 'true'})

# Read Lance dataset back using metadata catalog
ray_dataset = read_lance(namespace=namespace, table_id=["lance_minio_catalog", "schema", "my_table32"], storage_options = {'lance.storage.access_key_id' :  'x','lance.storage.endpoint' : 'http://minio:9000','lance.storage.secret_access_key' : 'x','lance.storage.allow_http' : 'true'})

# Perform distributed operations
result = ray_dataset.filter(lambda row: row["value"] < 100).count()
print(f"Filtered count: {result}")   

And then the logs message

>>> write_lance(data, namespace=namespace, table_id=["lance_minio_catalog", "schema","my_table32"], storage_options = {'lance.storage.access_key_id' :  'xx','lance.storage.endpoint' : 'http://minio:9000','lance.storage.secret_access_key' : 'xxx','lance.storage.allow_http' : 'true'})
2026-01-06 15:19:28,163	INFO streaming_executor.py:108 -- Starting execution of Dataset. Full logs are in /tmp/ray/session_2026-01-05_16-05-53_740074_85505/logs/ray-data
2026-01-06 15:19:28,164	INFO streaming_executor.py:109 -- Execution plan of Dataset: InputDataBuffer[Input] -> TaskPoolMapOperator[ReadRange->Map(<lambda>)->Write]
(ReadRange->Map(<lambda>)->Write pid=85885) [2026-01-06T07:19:35Z WARN  lance::dataset::write::insert] No existing dataset at s3://bucket1/lance_minio_catalog/schema/my_table32/, it will be created [repeated 20x across cluster] (Ray deduplicates logs by default. Set RAY_DEDUP_LOGS=0 to disable log deduplication, or see https://docs.ray.io/en/master/ray-observability/user-guides/configure-logging.html#log-deduplication for more options.)
(ReadRange->Map(<lambda>)->Write pid=85886) [2026-01-06T07:19:41Z WARN  lance::dataset::write::insert] No existing dataset at s3://bucket1/lance_minio_catalog/schema/my_table32/, it will be created [repeated 4x across cluster]
(ReadRange->Map(<lambda>)->Write pid=85887) [2026-01-06T07:19:46Z WARN  lance::dataset::write::insert] No existing dataset at s3://bucket1/lance_minio_catalog/schema/my_table32/, it will be created [repeated 17x across cluster]
(ReadRange->Map(<lambda>)->Write pid=85694) [2026-01-06T07:19:52Z WARN  lance::dataset::write::insert] No existing dataset at s3://bucket1/lance_minio_catalog/schema/my_table32/, it will be created [repeated 16x across cluster]
2026-01-06 15:19:53,973	INFO datasink.py:103 -- Write operation succeeded. Aggregated write results:
	- num_rows: 1000
	- size_bytes: 16000

>>> ray_dataset = read_lance(namespace=namespace, table_id=["lance_minio_catalog", "schema", "my_table32"], storage_options = {'lance.storage.access_key_id' :  'xx','lance.storage.endpoint' : 'http://minio:9000','lance.storage.secret_access_key' : 'xx','lance.storage.allow_http' : 'true'})
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/yuqi/project/lance-ray/lance_ray/io.py", line 114, in read_lance
    return read_datasource(
  File "/Users/yuqi/.pyenv/versions/3.10.13/lib/python3.10/site-packages/ray/_private/auto_init_hook.py", line 21, in auto_init_wrapper
    return fn(*args, **kwargs)
  File "/Users/yuqi/.pyenv/versions/3.10.13/lib/python3.10/site-packages/ray/data/read_api.py", line 387, in read_datasource
    requested_parallelism, _, inmemory_size = _autodetect_parallelism(
  File "/Users/yuqi/.pyenv/versions/3.10.13/lib/python3.10/site-packages/ray/data/_internal/util.py", line 176, in _autodetect_parallelism
    mem_size = datasource_or_legacy_reader.estimate_inmemory_data_size()
  File "/Users/yuqi/project/lance-ray/lance_ray/datasource.py", line 150, in estimate_inmemory_data_size
    if not self.fragments:
  File "/Users/yuqi/project/lance-ray/lance_ray/datasource.py", line 85, in fragments
    self._fragments = self.lance_dataset.get_fragments() or []
  File "/Users/yuqi/project/lance-ray/lance_ray/datasource.py", line 79, in lance_dataset
    self._lance_ds = lance.dataset(**dataset_options)
  File "/Users/yuqi/.pyenv/versions/3.10.13/lib/python3.10/site-packages/lance/__init__.py", line 237, in dataset
    ds = LanceDataset(
  File "/Users/yuqi/.pyenv/versions/3.10.13/lib/python3.10/site-packages/lance/dataset.py", line 443, in __init__
    self._ds = _Dataset(
ValueError: Dataset at path lance_minio_catalog/schema/my_table32 was not found: LanceError(IO): Generic N/A error: Encountered internal error. Please file a bug report at https://github.com/lance-format/lance/issues. Failed to get AWS credentials: CredentialsNotLoaded(CredentialsNotLoaded { source: Some("no providers in chain provided credentials") }), /Users/runner/work/lance/lance/rust/lance-io/src/object_store/providers/aws.rs:439:31, /Users/runner/work/lance/lance/rust/lance-io/src/object_store.rs:670:92, /Users/runner/work/lance/lance/rust/lance/src/dataset/builder.rs:628:35

Based on the logs, it seems that we have written data to the Lance table, but it fails to load it, however, I checked the location in MinIO, the location does not exists.

Start with lance_ray/io.py around read_lance and datasource.py around the fragments and lance_dataset properties shown in the traceback. Reproduce the write-then-read example with the MinIO storage options and inspect how the dataset path and options are passed to Lance. Done means read_lance can locate and load the dataset written by write_lance, with coverage for this MinIO configuration.