NVIDIA / NVIDIA/cudf

[FEA] No-op/perform element selection for (cum)max/min in cudf_polars if `Column.is_sorted`

Open
#21,914 0 comments 0 reactions 0 assignees View on GitHub
cudf-polars feature request
Dominant language
C++
Stars
9.8k
Forks
1.1k
Avg merge
3d 6m
Merged PRs (30d)
278

Description

(Identified by Claude)

* For cum_min/max, if the `Column.is_sorted` with an aligned `Column.order` relative to the min/max, the result is just the column itself and a scan is not needed

https://github.com/rapidsai/cudf/blob/01c7adae10d12da7a175fb616300f8190522b6e0/python/cudf_polars/cudf_polars/dsl/expressions/unary.py#L595-L598

* For min/max, if the `Column.is_sorted` and the data does not have `nan`s, the min/max is just the `plc.copying.get_element` of the first/last non-null element

https://github.com/rapidsai/cudf/blob/01c7adae10d12da7a175fb616300f8190522b6e0/python/cudf_polars/cudf_polars/dsl/expressions/aggregation.py#L217-L251

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.