Mark JSON-bearing string fields with is_json metadata

[adriangb]

Summary

• json_get_array returns List<Utf8> whose items are raw JSON strings, but the inner field was missing the is_json: true field metadata. This adds it on both the declared return type and the produced ListArray (via ListBuilder::with_field), so downstream consumers can detect that the items are JSON.
• json_get_json returns Utf8 containing a raw JSON sub-document; added is_json: true on the top-level return field via return_field_from_args.
• Centralized the metadata into a shared is_json_metadata() helper in common.rs and reused it in common_union.rs (was duplicated inline).

Audit notes

While auditing, I checked every UDF for missing JSON metadata:

UDF	Returns JSON?	Status
json_get_array	Inner Utf8 items are JSON	Fixed
json_get_json	Top-level Utf8 is raw JSON	Fixed
json_get / json_from_scalar	JsonUnion (array/object members are JSON)	Already correct via union_fields()
json_as_text	Mixed — strings are unwrapped, only objects/arrays return JSON	Intentionally not marked (would be misleading)
json_object_keys	List of object keys (plain strings)	N/A
json_get_str/int/float/bool, json_length, json_contains	Primitives	N/A

json_as_text is intentionally left unmarked because for Peek::String it returns the unescaped string value (no surrounding quotes), so the output is not consistently valid JSON. Happy to revisit if you'd prefer it marked.

Test plan

• cargo test — all 154 tests pass
• cargo clippy --all-targets — clean
• New test test_json_get_array_inner_field_is_json_metadata verifies the inner field metadata both in schema and in the produced ListArray
• New test test_json_get_json_is_json_metadata verifies the top-level field metadata

🤖 Generated with Claude Code

[Copilot]

Pull request overview

This PR standardizes how JSON-bearing string outputs are annotated across the JSON UDFs by ensuring Arrow Field metadata includes is_json: true wherever the returned Utf8 (or inner Utf8 list items) contains raw JSON, enabling downstream consumers to reliably detect JSON strings.

Changes:

• Add a shared is_json_metadata() helper and reuse it for JSON-bearing Field metadata.
• Mark json_get_array’s inner Utf8 list item field as JSON in both the declared return type and the produced ListArray builder field.
• Mark json_get_json’s returned Utf8 field as JSON via return_field_from_args, and add regression tests for both UDFs.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
src/common.rs	Introduces is_json_metadata() helper for consistent is_json field metadata creation.
src/common_union.rs	Reuses is_json_metadata() instead of duplicating inline metadata construction for union fields.
src/json_get_array.rs	Ensures the list item Field for json_get_array is marked is_json and that produced ListArrays carry the same metadata.
src/json_get_json.rs	Sets is_json metadata on the returned field via return_field_from_args.
tests/main.rs	Adds tests asserting JSON metadata is present on both schema fields and produced arrays.

---

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