lance-format / lance-format/lance
Add more sophistication to index choice when multiple indexes apply
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 7.1k
- Forks
- 852
- Avg merge
- 3d 18h
- Merged PRs (30d)
- 272
Description
Currently if a column has multiple indexes that could apply then we just choose the first index in the manifest. However, we should at the very least, prefer exact indexes over inexact indexes. Or maybe prefer up-to-date indexes over older indexes.
A very common pattern users might adopt is to create cheap zone map indexes (or bloom filter indexes) on most / all columns and then create the more intensive indexes later based on query patterns. Currently, if a user did this, we would continue using the zone map index, forcing the user to manually delete the zone map index when they trained the newer 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 locating the manifest logic that chooses among multiple applicable indexes. Compare the requested preference for exact indexes with the alternative of preferring newer indexes, then define and implement a deterministic choice that avoids selecting an older or less precise index when a better one is available.
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
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 30/100