lance-format / lance-format/lance
Thread SpillStore through all index builds + per-index codecs
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 7.1k
- Forks
- 852
- Avg merge
- 3d 18h
- Merged PRs (30d)
- 272
Description
With the SpillStore trait (#7300) and MemoryPool threading (#7301) in place, wire each index builder to spill intermediate state through SpillStore when it hits memory pressure.
Add thin per-index codecs beside each builder that (de)serialize intermediate state — posting-list segments, shuffle runs, btree pages — to scratch. The Arrow-vs-raw-bytes choice lives in these codecs.
This is the umbrella issue. Per-family work:
- FTS / inverted — #7307
- Vector IVF — #7308
- Scalar btree/bitmap — #7309
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 reading the SpillStore trait from #7300 and MemoryPool threading from #7301, then review the per-family issues #7307, #7308, and #7309. Determine how each index builder should spill posting-list segments, shuffle runs, or btree pages and where its codec belongs. Done means all listed index families use SpillStore under memory pressure with per-index serialization choices defined.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust
- Domain
- databases
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100