Implementation:Apache Paimon Ray Init

Knowledge Sources	Apache Paimon Ray Init
Domains	Data_Lake, Distributed_Computing
Last Updated	2026-02-07 00:00 GMT

Overview

External tool for initializing the Ray distributed computing framework for Paimon integration.

Description

ray.init() starts or connects to a Ray cluster runtime. When used with Paimon, it enables distributed table reads via to_ray() and distributed writes via write_ray(). Supports both local mode (single machine with multiple CPUs) and cluster mode (multi-node deployment).

Usage

Call ray.init() once at the beginning of any script that uses Paimon distributed operations. Use ignore_reinit_error=True for idempotent initialization in notebooks or scripts that may be re-executed.

Code Reference

Source Location

External Tool - ray.init() documentation

Signature

ray.init(
    address: Optional[str] = None,
    *,
    num_cpus: Optional[int] = None,
    num_gpus: Optional[int] = None,
    ignore_reinit_error: bool = False,
    **kwargs
) -> None

Import

import ray

I/O Contract

Inputs

Name	Type	Required	Description
address	Optional[str]	No	Ray cluster address; None for local mode
num_cpus	Optional[int]	No	Number of CPUs for local cluster
num_gpus	Optional[int]	No	Number of GPUs for local cluster
ignore_reinit_error	bool	No	Skip if already initialized (default False)

Outputs

Name	Type	Description
(return)	None	Initializes Ray runtime as a side effect

Usage Examples

Basic Usage

import ray

# Initialize local Ray cluster with 4 CPUs
ray.init(ignore_reinit_error=True, num_cpus=4)

Related Pages

Implements Principle

• Principle:Apache_Paimon_Ray_Cluster_Initialization

Requires Environment

• Environment:Apache_Paimon_Optional_Extensions