Skip to content

DataFusion Bloom v0.1.0

Choose a tag to compare

View all tags
@YimingQiao YimingQiao released this 17 Aug 02:31
· 16 commits to main since this release
v0.1.0
c7ae7b4

First public preview of Bloom for Apache DataFusion.

Highlights:

  • Adds Bloom transfer and materialization ahead of DataFusion native inner equi-joins.
  • Keeps FullRows as the default handoff and RowLocations as an experimental opt-in path.
  • Uses single-flight prepared samples with bounded session reuse.
  • Accounts retained handoffs, samples, and membership structures in DataFusion's memory pool.
  • Falls back to the untouched native plan when recoverable transfer work cannot be completed.
  • Includes reproducible JOB, CEB IMDB, STATS-CEB, and TPC-H benchmark tooling.

Compatibility: Apache DataFusion 54.x from 54.1.0, Rust 1.88 or newer. Licensed under MIT.

Assets 2
Loading