datafusion-contrib / datafusion-contrib/liquid-cache

[Epic] FSSTView in LiquidCache

Open
#301 0 comments 1 reaction 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Rust
Stars
452
Forks
51
Avg merge
3h 10m
Merged PRs (30d)
12

Description

I'm cooking some new string representation that can potentially increase many query's performance by orders of magnitude.

Design doc can be find here: https://github.com/XiangpengHao/liquid-cache/blob/main/dev/design/00-fsst-view.md

#297 has implemented most of the logic, #299 has added benchmark to show that it performs particularly well for predicates that can leverage string prefixes, e.g., eq, neq, lt, gt, etc.

Remaining work are:

  1. Make it the default string representation in LiquidCache. This requires some more rigorous testings and benchmarking to understand the trade-offs, especially that we need a bit more memory compare to the existing approach.
  2. Implement the disk eviction. The idea is to keep the prefix part in memory, and evict the fsst buffer to disk. And later use some FAST IO to read the fsst on-demand.

Contributor guide

No contributing guide indexed for this repository

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 design document at dev/design/00-fsst-view.md, then review issues #297 and #299 for the existing implementation and benchmarks. Done means establishing FSSTView as the default after rigorous testing and benchmarking, and implementing disk eviction for the FSST buffer with on-demand reads.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
databases, performance
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.