lance-format / lance-format/lance
Less RAM-intensive approach for 2.1+ blob reading
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 7.1k
- Forks
- 852
- Avg merge
- 3d 18h
- Merged PRs (30d)
- 272
Description
Currently the 2.1 blob scheduler loads all descriptions into RAM at initialize time. We could instead offer a 2-iops path where the descriptions are loaded on a per-page basis during schedule_ranges. This would introduce some indirect I/O but we can tackle that by spawning a task similar to how we handle full-zip variable data with repetition index.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by tracing the 2.1 blob scheduler's initialization and schedule_ranges paths, then compare them with the full-zip variable-data handling and repetition-index task mentioned in the issue. Done should provide a per-page description-loading path with bounded memory while accounting for the resulting indirect I/O.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- data-engineering, performance
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100