Skip to content

GH-50295: [C++][R] #include <ranges> in vector_select_k.cc breaks macOS CRAN and wasm builds #50297

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

Merged
thisisnic merged 4 commits into from
Jul 1, 2026
+331 −53
Merged

GH-50295: [C++][R] #include <ranges> in vector_select_k.cc breaks macOS CRAN and wasm builds #50297

Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
21b0c22
Don't use ranges
thisisnic Jun 30, 2026
05d44a5
Fix underflow
thisisnic Jun 30, 2026
f487731
Add tests for columns without non-null values
pitrou Jun 30, 2026
5c1f84a
Merge pull request #34 from pitrou/gh50295-additional-tests
thisisnic Jun 30, 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
18 changes: 8 additions & 10 deletions cpp/src/arrow/compute/kernels/vector_select_k.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,6 @@

#include <algorithm>
#include <queue>
#include <ranges>
#include <span>

#include "arrow/compute/function.h"
Expand Down Expand Up @@ -118,22 +117,22 @@ void HeapSortNonNullsToOutput(std::span<uint64_t> non_null_input_range, Comparat
return;
}
std::span<uint64_t> heap = non_null_input_range.subspan(0, output_range.size());
std::ranges::make_heap(heap, cmp);
std::make_heap(heap.begin(), heap.end(), cmp);

std::span<uint64_t> remaining_input = non_null_input_range.subspan(output_range.size());
for (uint64_t x_index : remaining_input) {
if (cmp(x_index, heap.front())) {
std::ranges::pop_heap(heap, cmp);
std::pop_heap(heap.begin(), heap.end(), cmp);
heap.back() = x_index;
std::ranges::push_heap(heap, cmp);
std::push_heap(heap.begin(), heap.end(), cmp);
}
}

// fill output in reverse when destructing,
// as the "worst" (next-to-would-have-been-replaced) element is at heap-top
for (auto& reverse_out_iter : std::ranges::reverse_view(output_range)) {
reverse_out_iter = heap.front(); // heap-top has the next element
std::ranges::pop_heap(heap, cmp);
for (int64_t i = output_range.size() - 1; i >= 0; --i) {

@pitrou pitrou Jun 30, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

Copilot didn't notice, but this has the same problem as below.

@thisisnic thisisnic Jun 30, 2026

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

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

Thanks!

output_range[i] = heap.front(); // heap-top has the next element
std::pop_heap(heap.begin(), heap.end(), cmp);
// Decrease heap-size by one
heap = heap.first(heap.size() - 1);
}
Expand Down Expand Up @@ -422,9 +421,8 @@ class ChunkedArraySelector : public TypeVisitor {
// so the heap must have been completely filled
DCHECK_EQ(heap.size(), output.non_null_like_range.size());

for (uint64_t& reverse_out_iter :
std::ranges::reverse_view(output.non_null_like_range)) {
reverse_out_iter =
for (int64_t i = output.non_null_like_range.size() - 1; i >= 0; --i) {
output.non_null_like_range[i] =

@pitrou pitrou Jun 30, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

The Copilot comment is right but the suggested solution is ugly. Could be something like:

Suggested change
for (int64_t i = output.non_null_like_range.size() - 1; i >= 0; --i) {
output.non_null_like_range[i] =
for (int64_t i = output.non_null_like_range.size(); i > 0; --i) {
output.non_null_like_range[i - 1] =

or:

Suggested change
for (int64_t i = output.non_null_like_range.size() - 1; i >= 0; --i) {
output.non_null_like_range[i] =
for (auto it = output.non_null_like_range.rbegin();
it != output.non_null_like_range.rend(); ++it) {
*it =

@pitrou pitrou Jun 30, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

I've pushed tests for the all-nulls case in thisisnic#34 @thisisnic

@thisisnic thisisnic Jun 30, 2026

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

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

Cheers @pitrou

heap.top().index + heap.top().offset; // heap-top has the next element
heap.pop();
}
Comment thread
thisisnic marked this conversation as resolved.
Outdated
Expand Down
Loading