Also emit canonical Arrow JSON extension metadata keys

[adriangb]

Summary

• Extend json_field_metadata() to emit Arrow's canonical JSON extension type keys (ARROW:extension:name = arrow.json, ARROW:extension:metadata = {}) alongside the existing is_json = true key.
• Re-export json_field_metadata at the crate root so external users (and our own tests) can stop duplicating the literal.
• Extend the two metadata-assertion tests added in Mark JSON-bearing string fields with is_json metadata #111 to verify all three keys, and route the test_json_get_utf8/test_json_get_large_utf8 fixtures through the helper.

Why

Marking JSON-bearing Utf8 fields with is_json is a convention private to this crate — no other Arrow tool recognizes it. Emitting the canonical keys as well makes those columns interoperable with the broader Arrow ecosystem (arrow-rs's Json extension type, pyarrow, DuckDB, Polars, …). The legacy is_json key is preserved so any downstream consumer that already keys off it keeps working.

The change is contained to one helper; all four production write sites (union_fields()'s array/object children, json_get_array's list item field, json_get_json's return field) already route through it. Detection via is_json_union is purely structural — comparing UnionFields against the OnceLock-cached union_fields() — so adding metadata keys to the cached value preserves equality on both sides.

Test plan

• cargo build
• cargo test — 154 passed, 0 failed (covers extended test_json_get_array_inner_field_json_metadata, test_json_get_json_json_metadata, fixture-only test_json_get_utf8 / test_json_get_large_utf8, and all is_json_union-gated paths)
• cargo clippy --all-targets -- -D warnings

🤖 Generated with Claude Code

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.05%. Comparing base (a3d9f62) to head (0b2e5e7).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #112      +/-   ##
==========================================
+ Coverage   82.99%   83.05%   +0.05%     
==========================================
  Files          17       17              
  Lines        1482     1487       +5     
  Branches     1482     1487       +5     
==========================================
+ Hits         1230     1235       +5     
  Misses        183      183              
  Partials       69       69              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

This PR updates the crate’s JSON-field metadata helper to emit Arrow’s canonical JSON extension metadata (for better interoperability across the Arrow ecosystem), while keeping the existing legacy is_json=true marker for backward compatibility. It also exposes the helper at the crate root and updates tests/fixtures to rely on it.

Changes:

• Extend json_field_metadata() to include ARROW:extension:name=arrow.json and ARROW:extension:metadata={} alongside legacy is_json=true.
• Re-export json_field_metadata from the crate root for external/test reuse.
• Update metadata assertion tests and route return-field fixtures through json_field_metadata().

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
src/common_union.rs	Expands JSON field metadata to include canonical Arrow extension keys while preserving is_json.
src/lib.rs	Re-exports json_field_metadata at the crate root.
tests/main.rs	Updates tests/fixtures to use and validate JSON field metadata.

Comments suppressed due to low confidence (1)

tests/main.rs:196

• assert_json_field_metadata currently only asserts the canonical Arrow extension keys, but this PR’s stated back-compat goal is to also preserve the legacy is_json = true key. As written, the tests would still pass if is_json were accidentally removed later. Please extend this helper (or compare against json_field_metadata()) so the tests verify all expected keys, including is_json.

fn assert_json_field_metadata(metadata: &HashMap<String, String>) {
    assert_eq!(
        metadata.get("ARROW:extension:name").map(String::as_str),
        Some("arrow.json")
    );
    assert_eq!(metadata.get("ARROW:extension:metadata").map(String::as_str), Some("{}"));
}

#[tokio::test]
async fn test_json_get_equals() {

---

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