NVIDIA / NVIDIA/cudf

[FEA] Support `LazyFrame.rolling()` for streaming cuDF-Polars

Open
#22,033 0 comments 0 reactions 0 assignees View on GitHub
cudf-polars feature request rapids-mpf
Dominant language
C++
Stars
9.8k
Forks
1.1k
Avg merge
3d 6m
Merged PRs (30d)
278

Description

We want `LazyFrame.rolling(...)` supported for streaming execution with the rapidsmpf runtime. Today Rolling IR falls back to single-partition execution, which is correct but not scalable.

## Scope

1. **Single-rank**: Implement a rolling actor path for `LazyFrame.rolling(...)`.
2. **Multi-rank**: Explicitly depend on a HaloExchange API in RapidsMPF (E.g. https://github.com/rapidsai/rapidsmpf/pull/947)

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.