GH-47103: [Statistics][C++] Implement Statistics specification attribute ARROW:null_count:approximate

[andishgar]

Rationale for this change

Enable ARROW:null_count:approximate support for arrow::ArrayStatistics, along with the corresponding Python binding.

What changes are included in this PR?

Enable ARROW:null_count:approximate in C++ and bind it to ArrayStatistics in Python.

Are these changes tested?

Yes, I ran the relevant unit tests.

Are there any user-facing changes?

Yes. The type of arrow::ArrayStatistics::null_count has been changed from std::optional<int64_t> to std::optional<CountType>, and a new field is_null_count_exact has been added to ArrayStatistics in Python.

• GitHub Issue: [Statistics][C++] Implement Statistics specification attribute ARROW:null_count:approximate #47103

[github-actions]

⚠️ GitHub issue #47103 has been automatically assigned in GitHub to PR creator.

[andishgar]

@kou
Regarding the Ruby binding, would it be possible to ask someone to work on it?

[kou]

Yes. I'll do it.

[Copilot]

Pull Request Overview

This PR adds support for approximate null counts in Arrow array statistics by extending the null_count field to support both exact (int64_t) and approximate (double) values using std::variant. This aligns with the existing pattern used for distinct_count.

Key changes:

• Changed null_count from std::optional<int64_t> to std::optional<CountType> (variant of int64_t and double)
• Added is_null_count_exact property to distinguish between exact and approximate null counts
• Updated all related tests and comparison logic to handle the variant type

Reviewed Changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
cpp/src/arrow/array/statistics.h	Changed null_count type from int64_t to CountType variant
cpp/src/arrow/record_batch.cc	Added logic to handle both exact and approximate null counts when creating statistics arrays
cpp/src/arrow/compare.cc	Updated equality comparison to use ArrayStatisticsOptionalValueEquals for null_count
cpp/src/arrow/record_batch_test.cc	Renamed test and added new test for approximate null count
cpp/src/arrow/array/statistics_test.cc	Added test for approximate null count and updated existing tests
cpp/src/arrow/array/array_test.cc	Updated variable type and test assertions to handle variant null_count
cpp/src/parquet/arrow/arrow_statistics_test.cc	Updated assertions to extract int64_t from variant null_count
python/pyarrow/includes/libarrow.pxd	Changed null_count type from int64_t to CArrayStatisticsCountType
python/pyarrow/array.pxi	Added is_null_count_exact property and updated null_count to handle variant
python/pyarrow/tests/parquet/test_parquet_file.py	Added assertion to verify is_null_count_exact is True

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.