Eventual-Inc/Daft - CodSpeed

Daft

Blog Docs Changelog

Performance History

Latest Results

chore(ci): retrigger CI (flaky style + integration-test-io docker infra)

hello-peter-tang:perf/6340-lm-pipelined-predicate-eval

2 hours ago

chore(ci): retrigger CI (flaky integration-test-io docker digest-mismatch)

hello-peter-tang:startswith-stats-pruning

2 hours ago

test(optimizer): make starts_with pushdown assertions order-independent The empty-prefix and projection cases scan both row groups and compared `to_pydict()` against a hard-coded row order. A parallel (ray) runner does not guarantee the relative order of row groups, so these assertions were flaky -- the ray unit-test job saw the two row groups reversed. Compare the produced values order-independently (sorted / Counter) instead. The pruned cases still assert exact lists since they read a single row group.

hello-peter-tang:startswith-stats-pruning

3 hours ago

refactor(parquet): bundle prefilter inputs into a context struct Replace the three `#[allow(clippy::too_many_arguments)]` suppressions on `prefilter_rg_monolithic`/`prefilter_rg_pipelined`/`build_rg_inputs` with a shared `RgPrefilterCtx` borrowing struct, so the helpers stay under the lint threshold without suppressing it. Also tidy the reader tests: hoist the per-test inline `use` blocks to module scope and assert the monolithic-fallback cases actually take the monolithic path via `assert_prefilter_strategy`.

hello-peter-tang:perf/6340-lm-pipelined-predicate-eval

3 hours ago

Merge remote-tracking branch 'upstream/main' into iceberg-partition-repr-parity

jackylee-ch:iceberg-partition-repr-parity

6 hours ago

style: fix codespell/ruff lint in starts_with rewrite tests - add codespell:ignore for the 'abd' successor-prefix literals in the Rust rule tests (increment('abc') == 'abd', flagged as a typo for 'and'/'bad') - rewrite the pushdown test module docstring to satisfy ruff D205 - mark the daft.sql-captured df with noqa: F841 and collapse a dict literal to match ruff-format

hello-peter-tang:startswith-stats-pruning

10 hours ago

rchowell/jev

19 hours ago

rchowell/jev

22 hours ago

Latest Branches

0%

perf(parquet): LM-pipelined predicate evaluation with stats-based group ordering#7488

3 hours ago

fa19d54

hello-peter-tang:perf/6340-lm-pipelined-predicate-eval

0%

perf(optimizer): rewrite starts_with filters into range predicates for pushdown pruning#7441

3 hours ago

9869e8f

hello-peter-tang:startswith-stats-pruning

0%

fix(iceberg): support timezone-aware identity partition writes#7443

6 hours ago

12fbb43

jackylee-ch:iceberg-partition-repr-parity

© 2026 CodSpeed Technology

Home Terms Privacy Docs