apache / apache/datafusion

[DISCUSSION] Add `DedicatedExecutor` into the DataFusion crates to make using multiple threadpools easier

Open
#14,285 0 comments 1 reaction 0 assignees View on GitHub
enhancement
Dominant language
Rust
Stars
9.3k
Forks
2.4k
Avg merge
3d 7h
Merged PRs (30d)
344

Description

### Is your feature request related to a problem or challenge?

- Related to https://github.com/apache/datafusion/issues/12393

### Describe the solution you'd like

As explained on the following PR:
- https://github.com/apache/datafusion/pull/14286

I implemented a version of `DedicatedExecutor` and an example showing how to use multiple thread pools with DataFusion. However, the example is currently very large as it uses an abstraction called `DedicatedExecutor` to manage sending work back and forth between the runtimes

If users want to do the same thing, they will need to copy `DedicatedExecutor` into their code (or make something similar).

### Describe alternatives you've considered

If `DedicatedExecutor` is a useful API for multiple people I think we should consider moving it into one of the core crates

### Additional context

_No response_

Contributor guide

Open the contributing guide

Research direction

Start by reading the related issue 12393 and pull request 14286, which contain the current DedicatedExecutor implementation and example. Determine whether the abstraction belongs in a DataFusion core crate and what API maintainers want; the work is done when the crate placement and public API are agreed and implemented with appropriate coverage.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
backend
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.