Skip to content

ranges: restore safe join selection and residual interval filtering#34

Open
mwiewior wants to merge 2 commits into
masterfrom
perf/issue-33-interval-join
Open

ranges: restore safe join selection and residual interval filtering#34
mwiewior wants to merge 2 commits into
masterfrom
perf/issue-33-interval-join

Conversation

@mwiewior
Copy link
Copy Markdown
Contributor

Summary

  • restore DataFusion join_selection for bio sessions without letting it swap side-sensitive interval joins
  • split the interval rewrite into a pre-join-selection pass for CoitreesNearest and CoitreesCountOverlaps, and a post-join-selection pass for symmetric inner overlap joins
  • extract residual conjuncts from interval join filters and apply them after interval candidate generation inside IntervalJoinExec
  • normalize interval bounds to Int32 once, use the computed build-side bucket estimate, and reuse the interval join scratch buffers and output helpers
  • add coverage for residual-filter rewrites and for non-symmetric methods so join selection cannot silently change their semantics

Testing

  • cargo test -p datafusion-bio-function-ranges

Partially addresses #33.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant