mirror of
https://github.com/lancedb/lancedb.git
synced 2026-08-18 12:08:35 +00:00
sanskar singh bhardwaj
7dfdfe6401
## Summary Fixes #2339. `merge_insert()` on the remote client could mask the real cause of a mid-stream input error, reporting only: > stream error sent by user: unexpected internal error ## Root cause There were two divergent streaming-write code paths in the remote client: - `add()` uses `RemoteInsertExec`, which streams the request body through a `tokio::sync::oneshot` error side-channel and drains it before reporting the HTTP result. If the input stream errors mid-body, the original error is recovered. - `merge_insert()` used a legacy path (`send_streaming` -> `reader_as_body`) that piped arrow `Some(Err(e))` straight into the HTTP2 request body. Hyper swallows body-stream errors under HTTP2 (see hyperium/hyper#2547), so the original error was lost and only the generic transport error surfaced. ## Fix Consolidate both write paths onto the side-channel mechanism instead of patching the legacy path: - Generalize `RemoteInsertExec` into `RemoteWriteExec`, carrying a `WriteOp` enum (`Insert { overwrite }` | `MergeInsert { query, timeout }`) that selects the endpoint, query params, request-timeout header, and response parsing. The executor returns a `WriteResult` enum (`Add` | `Merge`) with typed accessors, and `with_new_children` still resets the result so the rescannable retry loop is unaffected. - Route `merge_insert()` through `RemoteWriteExec`. The public API only accepts a `RecordBatchReader` (not rescannable), so the reader is buffered into a `Vec<RecordBatch>` before the retry loop to preserve the previous retry-on-retryable-status behaviour. This mirrors what the old `send_streaming(with_retry=true)` path already did. - Remove the now-unused `send_streaming` / `reader_as_body` / `buffer_reader` / `make_reader` helpers. Multipart stays insert-only (the server has no multipart merge_insert endpoint), so that hot path is behaviorally unchanged. ## Testing - Added `test_merge_insert_input_error_surfaces_original`, which drives an erroring input through the single-request `merge_insert` path and asserts the original error (`boom`) is surfaced rather than the masked HTTP error. Confirmed it fails without the side-channel drain (it then reports a masked `500 ... request or response body error`). - Full suite green: `cargo test -p lancedb --lib --features remote` -> 694 passed, 0 failed. Includes the existing `test_merge_insert_retries_on_409`, confirming retry behaviour is preserved.
The Multimodal AI Lakehouse
How to Install ✦ Detailed Documentation ✦ Tutorials and Recipes ✦ Contributors
The ultimate multimodal data platform for AI/ML applications.
LanceDB is designed for fast, scalable, and production-ready vector search. It is built on top of the Lance columnar format. You can store, index, and search over petabytes of multimodal data and vectors with ease. LanceDB is a central location where developers can build, train and analyze their AI workloads.
Demo: Multimodal Search by Keyword, Vector or with SQL
Star LanceDB to get updates!
⭐ Click here ⭐ to see how fast we're growing!
Key Features:
- Fast Vector Search: Search billions of vectors in milliseconds with state-of-the-art indexing.
- Comprehensive Search: Support for vector similarity search, full-text search and SQL.
- Multimodal Support: Store, query and filter vectors, metadata and multimodal data (text, images, videos, point clouds, and more).
- Advanced Features: Zero-copy, automatic versioning, manage versions of your data without needing extra infrastructure. GPU support in building vector index.
Products:
- Open Source & Local: 100% open source, runs locally or in your cloud. No vendor lock-in.
- Cloud and Enterprise: Production-scale vector search with no servers to manage. Complete data sovereignty and security.
Ecosystem:
- Columnar Storage: Built on the Lance columnar format for efficient storage and analytics.
- Seamless Integration: Python, Node.js, Rust, and REST APIs for easy integration. Native Python and Javascript/Typescript support.
- Rich Ecosystem: Integrations with LangChain 🦜️🔗, LlamaIndex 🦙, Apache-Arrow, Pandas, Polars, DuckDB and more on the way.
How to Install:
Follow the Quickstart doc to set up LanceDB locally.
API & SDK: We also support Python, Typescript and Rust SDKs
| Interface | Documentation |
|---|---|
| Python SDK | https://lancedb.github.io/lancedb/python/python/ |
| Typescript SDK | https://lancedb.github.io/lancedb/js/globals/ |
| Rust SDK | https://docs.rs/lancedb/latest/lancedb/index.html |
| REST API | https://docs.lancedb.com/api-reference/rest |
Join Us and Contribute
We welcome contributions from everyone! Whether you're a developer, researcher, or just someone who wants to help out.
If you have any suggestions or feature requests, please feel free to open an issue on GitHub or discuss it on our Discord server.
Check out the GitHub Issues if you would like to work on the features that are planned for the future. If you have any suggestions or feature requests, please feel free to open an issue on GitHub.
Contributors
Stay in Touch With Us