Skip to content

perf: Avoid unnecessary rechunk when sorting already sorted DataFrame #27264

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
gab23r wants to merge 2 commits into
base: main
Choose a base branch
from
+11 −2
Open

perf: Avoid unnecessary rechunk when sorting already sorted DataFrame #27264

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
a72bd97
perf: Avoid unnecessary rechunk when sorting already sorted DataFrame
Apr 10, 2026
4914e71
add test
Apr 10, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 1 addition & 2 deletions crates/polars-mem-engine/src/executors/sort.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,10 +13,9 @@ impl SortExec {
fn execute_impl(
&mut self,
state: &ExecutionState,
mut df: DataFrame,
df: DataFrame,
) -> PolarsResult<DataFrame> {
state.should_stop()?;
df.rechunk_mut_par();

let height = df.height();

Expand Down
10 changes: 10 additions & 0 deletions py-polars/tests/unit/operations/test_sort.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1289,3 +1289,13 @@ def test_sort_by_empty_list_eval_25433() -> None:
out = df.select(pl.col.a.list.eval(pl.element().sort_by(pl.element())))
expected = pl.DataFrame({"a": [sorted(some_list), []]})
assert_frame_equal(out, expected)


def test_sort_already_sorted_no_rechunk_25733() -> None:
df1 = pl.DataFrame({"a": [1, 2], "b": [3, 4]})
df2 = pl.DataFrame({"a": [3, 4], "b": [5, 6]})
df = pl.concat([df1, df2], rechunk=False).with_columns(pl.col("a").set_sorted())
assert df.n_chunks() == 2

result = df.sort("a")
assert result.n_chunks() == 2 # No rechunk happened
Loading