Skip to content

compatible issue related with ray==2.50.0 #50

@chenghao-guo

Description

@chenghao-guo

Issue:
Feedback from lance Chinese community:

import ray
from lance_ray import read_lance, write_lance

# Initialize Ray
ray.init()

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

# Write to Lance format
write_lance(data, "my_dataset.lance")

# Read Lance dataset back as Ray dataset
ray_dataset = read_lance("my_dataset.lance")

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

Failure in

TypeError: LanceDatasource.get_read_tasks() got an unexpected keyword argument 'per_task_row_limit'

Cause:
ray-project/ray#55239 added the parameter and caused this

Workaround
pip uninstall ray
pip install ray==2.49.0

Issue tracking:
I will take a further look later and make lance-ray compatible with ray==2.50.0

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions