Skip to content
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

Jump to bottom

fix: Dropped/shifted rows in parquet scan with streaming=True #18766

Merged
merged 2 commits into from
Sep 17, 2024
Merged

fix: Dropped/shifted rows in parquet scan with streaming=True #18766

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
f52cb09
fix: Dropped/shifted rows in parquet scan with `streaming=True`
nameexhaustion Sep 16, 2024
31a7dc8
Python Polars 1.7.2
nameexhaustion Sep 16, 2024
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
2 changes: 1 addition & 1 deletion Cargo.lock
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

7 changes: 2 additions & 5 deletions crates/polars-io/src/parquet/read/read_impl.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -644,18 +644,15 @@ fn rg_to_dfs_par_over_rg(
continue;
}

row_groups.push((i, rg_md, rg_slice, row_count_start));
row_groups.push((rg_md, rg_slice, row_count_start));
}

let dfs = POOL.install(|| {
// Set partitioned fields to prevent quadratic behavior.
// Ensure all row groups are partitioned.
row_groups
.into_par_iter()
.enumerate()
.map(|(iter_idx, (_rg_idx, _md, slice, row_count_start))| {
let md = &file_metadata.row_groups[iter_idx];

.map(|(md, slice, row_count_start)| {
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

iter_idx was an index into the previous version of partitioned metadata - I was just meant to restore this to directly use md from par_iter()

if slice.1 == 0 || use_statistics && !read_this_row_group(predicate, md, schema)? {
return Ok(None);
}
Expand Down
2 changes: 1 addition & 1 deletion py-polars/Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
[package]
name = "py-polars"
version = "1.7.1"
version = "1.7.2"
edition = "2021"

[lib]
Expand Down
12 changes: 12 additions & 0 deletions py-polars/tests/unit/io/test_lazy_parquet.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -529,3 +529,15 @@ def trim_to_metadata(path: str | Path) -> None:
assert_frame_equal(
pl.scan_parquet(path).slice(-1, (1 << 32) - 1).collect(), df.tail(1)
)


@pytest.mark.parametrize("streaming", [True, False])
def test_parquet_row_groups_shift_bug_18739(tmp_path: Path, streaming: bool) -> None:
tmp_path.mkdir(exist_ok=True)
path = tmp_path / "data.bin"

df = pl.DataFrame({"id": range(100)})
df.write_parquet(path, row_group_size=1)

lf = pl.scan_parquet(path)
assert_frame_equal(df, lf.collect(streaming=streaming))