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 ignored limit on lexsort_to_indices #2991

Merged
merged 3 commits into from
Oct 31, 2022
Merged

Fix ignored limit on lexsort_to_indices #2991

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
a2205af
Fix ignored limit on lexsort_to_indices
alamb Oct 31, 2022
8afe667
Update comments
alamb Oct 31, 2022
2b44518
Update arrow/src/compute/kernels/sort.rs
alamb Oct 31, 2022
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
40 changes: 33 additions & 7 deletions arrow/src/compute/kernels/sort.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -950,7 +950,7 @@ pub fn lexsort_to_indices(
});

Ok(UInt32Array::from_iter_values(
value_indices.iter().map(|i| *i as u32),
value_indices.iter().take(len).map(|i| *i as u32),
Copy link
Contributor Author

Choose a reason for hiding this comment

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

this is the bugfix

))
}

Expand Down Expand Up @@ -1422,6 +1422,18 @@ mod tests {
}
}

/// slice all arrays in expected_output to offset/length
fn slice_arrays(
expected_output: Vec<ArrayRef>,
offset: usize,
length: usize,
) -> Vec<ArrayRef> {
expected_output
.into_iter()
.map(|array| array.slice(offset, length))
.collect()
}

fn test_sort_binary_arrays(
data: Vec<Option<Vec<u8>>>,
options: Option<SortOptions>,
Expand Down Expand Up @@ -3439,8 +3451,10 @@ mod tests {
Some(2),
Some(17),
])) as ArrayRef];
test_lex_sort_arrays(input.clone(), expected, None);
test_lex_sort_arrays(input.clone(), expected.clone(), None);
test_lex_sort_arrays(input.clone(), slice_arrays(expected, 0, 2), Some(2));
Copy link
Contributor Author

Choose a reason for hiding this comment

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

The only place on master that a limit is passed to lexsort_to_indices in the tests is immediately below here. However, very sadly, there is a special case code path for single arrays that doesn't hit the bug path

        let expected = vec![Arc::new(PrimitiveArray::<Int64Type>::from(vec![
            Some(-1),
            Some(0),
            Some(2),
        ])) as ArrayRef];
        test_lex_sort_arrays(input, expected, Some(3));

Copy link
Contributor Author

Choose a reason for hiding this comment

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

This addition is strictly unnecessary from a coverage perspective (it was already covered), but I wanted to make the test_lex_sort_arrays based tests all consistently patterned so it was easier to reason about coverage


// Explicitly test a limit on the sort as a demonstration
let expected = vec![Arc::new(PrimitiveArray::<Int64Type>::from(vec![
Some(-1),
Some(0),
Expand Down Expand Up @@ -3519,7 +3533,8 @@ mod tests {
Some(-2),
])) as ArrayRef,
];
test_lex_sort_arrays(input, expected, None);
test_lex_sort_arrays(input.clone(), expected.clone(), None);
test_lex_sort_arrays(input, slice_arrays(expected, 0, 2), Some(2));
Copy link
Contributor Author

Choose a reason for hiding this comment

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

This test fails immediately without the fix -- the output is too big!)


// test mix of string and in64 with option
let input = vec![
Expand Down Expand Up @@ -3562,7 +3577,8 @@ mod tests {
Some("7"),
])) as ArrayRef,
];
test_lex_sort_arrays(input, expected, None);
test_lex_sort_arrays(input.clone(), expected.clone(), None);
test_lex_sort_arrays(input, slice_arrays(expected, 0, 3), Some(3));

// test sort with nulls first
let input = vec![
Expand Down Expand Up @@ -3605,7 +3621,8 @@ mod tests {
Some("world"),
])) as ArrayRef,
];
test_lex_sort_arrays(input, expected, None);
test_lex_sort_arrays(input.clone(), expected.clone(), None);
test_lex_sort_arrays(input, slice_arrays(expected, 0, 1), Some(1));

// test sort with nulls last
let input = vec![
Expand Down Expand Up @@ -3648,7 +3665,8 @@ mod tests {
None,
])) as ArrayRef,
];
test_lex_sort_arrays(input, expected, None);
test_lex_sort_arrays(input.clone(), expected.clone(), None);
test_lex_sort_arrays(input, slice_arrays(expected, 0, 2), Some(2));

// test sort with opposite options
let input = vec![
Expand Down Expand Up @@ -3695,7 +3713,15 @@ mod tests {
Some("foo"),
])) as ArrayRef,
];
test_lex_sort_arrays(input, expected, None);
test_lex_sort_arrays(input.clone(), expected.clone(), None);
test_lex_sort_arrays(
input.clone(),
slice_arrays(expected.clone(), 0, 5),
Some(5),
);

// Limiting by more rows than present is ok
test_lex_sort_arrays(input, slice_arrays(expected, 0, 5), Some(10));
}

#[test]
Expand Down