lance-format / lance-format/lance

Make _take_rows & _rowid Public

Open
#6,200 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement
Dominant language
Rust
Stars
7.1k
Forks
852
Avg merge
3d 18h
Merged PRs (30d)
272

Description

Hey Lance people, I'm currently using the scanner api to allow sql predicates to be passed to form a subset index of rows that match predicate. I'm saving off the _rowid from each of these as the lookup key. This is for selected pre-filtering on metadata light columns that is followed up by _take_rows on heavy columns across the dataset. I'm doing this to enable map style torch datasets where I can't rely on sql scanner predication through the scanner being only iterable.

I'm using these privates to do that work so I don't have to manually add a global offset index in my distributed write or parse out the fragment id and offset from the _rowid in my code. Is it possible to support these privates as public so there's a guarantee of the api not breaking?

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the scanner API entry points for the private _take_rows and _rowid methods, then trace how callers use them for predicate filtering and row lookup. Define what public API guarantee is required and verify that both operations are exposed without requiring callers to recreate global offsets or parse fragment IDs.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
api, data
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.