feat(bulk): attach auto-create table metadata to Arrow schema

[v0y4g3r]

Summary

Aligns the Rust bulk writer with the latest GreptimeDB Enterprise Flight bulk auto-create contract from GreptimeTeam/greptimedb-enterprise#845 (HEAD 9b277b5b).

The client now provides both parts required by the server:

• Strict Arrow field metadata describing the schema to create
• The Flight request hint that explicitly enables or disables automatic creation

API

Existing constructors remain source-compatible and default automatic table creation to disabled:

let writer = inserter
    .create_bulk_stream_writer(&schema, options)
    .await?;

Use the new constructor to enable it for a writer:

let writer = inserter
    .create_bulk_stream_writer_with_auto_create_table(&schema, options, true)
    .await?;

The client sends x-greptime-hints: auto_create_table=true|false. A true hint only takes effect when automatic table creation is also enabled globally on the server.

Changes

• Attach greptime:semantic_type (timestamp/tag/field) to every Arrow field
• Emit greptime:type: Json for JSON fields encoded as Arrow Binary, never on the timestamp field
• Reject schemas containing multiple timestamp columns before opening the stream
• Add source-compatible explicit auto-create constructors to BulkInserter and BulkStreamWriter; existing constructors send false
• Forward auto_create_table through the shared x-greptime-hints metadata while preserving database and authentication headers

Testing

• Verify semantic metadata values, allowed key cardinality, JSON/Binary/Decimal128 behavior, and multiple-timestamp rejection
• Verify exact true/false hint pairs
• Verify the constructed Flight request contains hint, database, and Basic authentication metadata
• cargo check --all-targets
• cargo clippy --all-targets -- -D warnings
• cargo fmt --check
• cargo test -- --skip integration (47 tests passed)

[Copilot]

Pull request overview

This PR extends the bulk insert Arrow schema to include GreptimeDB Enterprise auto-create-table contract metadata on every Arrow Field, ensuring the server can infer each column’s semantic role (timestamp/tag/field) and extended type information (e.g., JSON) during the first Flight bulk insert.

Changes:

• Introduces greptime:semantic_type field metadata (timestamp/tag/field) on all bulk-insert Arrow fields via a shared column_to_arrow_field helper.
• Adds greptime:type=Json metadata for JSON columns (encoded as Arrow Binary) while ensuring it is never set on timestamp columns.
• Adds early rejection for schemas containing multiple timestamp columns via a new MultipleTimestampColumns error, plus unit tests covering the new contract behavior.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
src/error.rs	Adds a new MultipleTimestampColumns error variant used to reject invalid schemas.
src/bulk.rs	Attaches GreptimeDB contract metadata to Arrow fields, adds multi-timestamp validation, and adds unit tests for the metadata contract and schema validation.

---

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