LLKV Table

Work in Progress

Columnar table using the LLKV toolkit.

This crate is designed to work directly with Arrow RecordBatch and does not provide any additional abstraction over the batch data model beyond how batches are queried and streamed. Data is fed into tables and retrieved from tables as batches of RecordBatch.

Purpose

• Provide schema-aware table abstraction built on llkv-column-map columnar storage.
• Manage table metadata via the system catalog (table 0).
• Handle MVCC columns (created_by, deleted_by) for transaction support.
• Support table scans with projection, filtering, and ordering.

Design Notes

• Tables use llkv-column-map for physical storage but add schema validation and field ID tracking.
• Integrates with llkv-transaction for row-level visibility filtering.
• Used by llkv-runtime for executing SQL operations and by llkv-executor for SELECT queries.

License

Licensed under the Apache-2.0 License.