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

GH-38770: [C++][Python] RecordBatch.filter() segfaults if passed a ChunkedArray #40971

Merged
merged 4 commits into from
May 8, 2024
Merged

GH-38770: [C++][Python] RecordBatch.filter() segfaults if passed a ChunkedArray #40971

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
6ae879c
Initial commit
AlenkaF Apr 3, 2024
c096240
Move the check to C++ and raise error
AlenkaF Apr 4, 2024
b4c730f
In case of chunked mask and RecordBatch use FilterTable
AlenkaF Apr 11, 2024
236d3bd
Move the change to FilterRecordBatch and concatenate the filter if ne…
AlenkaF Apr 15, 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
26 changes: 20 additions & 6 deletions cpp/src/arrow/compute/kernels/vector_selection_filter_internal.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,6 +22,7 @@
#include <type_traits>
#include <vector>

#include "arrow/array/concatenate.h"
#include "arrow/array/data.h"
#include "arrow/buffer_builder.h"
#include "arrow/chunked_array.h"
Expand Down Expand Up @@ -924,12 +925,26 @@ Result<std::shared_ptr<RecordBatch>> FilterRecordBatch(const RecordBatch& batch,
return Status::Invalid("Filter inputs must all be the same length");
}

// Convert filter to selection vector/indices and use Take
// Fetch filter
const auto& filter_opts = *static_cast<const FilterOptions*>(options);
ARROW_ASSIGN_OR_RAISE(
std::shared_ptr<ArrayData> indices,
GetTakeIndices(*filter.array(), filter_opts.null_selection_behavior,
ctx->memory_pool()));
ArrayData filter_array;
switch (filter.kind()) {
case Datum::ARRAY:
filter_array = *filter.array();
break;
case Datum::CHUNKED_ARRAY: {
ARROW_ASSIGN_OR_RAISE(auto combined, Concatenate(filter.chunked_array()->chunks()));
filter_array = *combined->data();
break;
}
default:
return Status::TypeError("Filter should be array-like");
}
AlenkaF marked this conversation as resolved.
Show resolved Hide resolved

// Convert filter to selection vector/indices and use Take
ARROW_ASSIGN_OR_RAISE(std::shared_ptr<ArrayData> indices,
GetTakeIndices(filter_array, filter_opts.null_selection_behavior,
ctx->memory_pool()));
std::vector<std::shared_ptr<Array>> columns(batch.num_columns());
for (int i = 0; i < batch.num_columns(); ++i) {
ARROW_ASSIGN_OR_RAISE(Datum out, Take(batch.column(i)->data(), Datum(indices),
Expand Down Expand Up @@ -1038,7 +1053,6 @@ class FilterMetaFunction : public MetaFunction {
}

if (args[0].kind() == Datum::RECORD_BATCH) {
auto values_batch = args[0].record_batch();
ARROW_ASSIGN_OR_RAISE(
std::shared_ptr<RecordBatch> out_batch,
FilterRecordBatch(*args[0].record_batch(), args[1], options, ctx));
Expand Down
5 changes: 5 additions & 0 deletions python/pyarrow/tests/test_compute.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1344,6 +1344,11 @@ def test_filter_record_batch():
expected = pa.record_batch([pa.array(["a", "e"])], names=["a'"])
assert result.equals(expected)

# GH-38770: mask is chunked array
chunked_mask = pa.chunked_array([[True, False], [None], [False, True]])
result = batch.filter(chunked_mask)
assert result.equals(expected)

result = batch.filter(mask, null_selection_behavior="emit_null")
expected = pa.record_batch([pa.array(["a", None, "e"])], names=["a'"])
assert result.equals(expected)
Expand Down
Loading