Eventual-Inc/Daft - CodSpeed

Daft

Blog Docs Changelog

Performance History

Latest Results

fix(deps): revert storage stack and azure-storage-blob bumps - gcsfs 2026.2.0 causes infinite recursion in `fs.ls()` for GCS integration tests - azure-storage-blob 12.28.0 uses API version 2026-02-06 which Azurite emulator doesn't support, breaking catalog integration tests - s3fs/boto3/adlfs/fsspec reverted to match (storage stack is tightly coupled through fsspec version pins) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

desmond/fix-dependabot-6598

4 hours ago

fix(deps): revert pylance bump - major version breaks distributed index API pylance 3.0.1 changed the `merge_index_metadata` API, breaking the lance distributed indexing tests. Revert to 0.39.0 until the integration code is updated. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

desmond/fix-dependabot-6598

4 hours ago

chore(deps): bump compatible dependencies from dependabot PR #6598 Cherry-picks the subset of bumps from dependabot that are compatible with the existing dependency graph and the exclude-newer constraint: - pylance 0.39.0 -> 3.0.1 (lance extra upper bound adjusted) - AWS stack: s3fs 2024.9.0 -> 2026.2.0, boto3 1.36.2 -> 1.42.70, moto 5.1.21 -> 5.1.22, adlfs 2024.7.0 -> 2026.2.0, azure-storage-blob 12.27.1 -> 12.28.0, gcsfs 2024.9.0 -> 2026.2.0 - clickhouse-connect 0.11.0 -> 0.14.1 - griffe 2.0.0 -> 2.0.2 - deltalake Windows pin 0.5.8 -> 1.5.0 - fsspec upper bound <2026.3.0 -> <2026.4.0 Reverted bumps that had cross-package conflicts: - vllm/torch/torchvision (tight coupling with huggingface-hub) - opencv-python (requires numpy>=2, but dev pins 1.26.4 for Python <3.13) - Pillow, aiohttp, google-genai, etc. (published after exclude-newer cutoff) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

desmond/fix-dependabot-6598

5 hours ago

fix(scan): skip getting bytes when range start equals end in daft async reader (#6602) ## Changes Made Check whether `range.start == range.end` in `RangesContainer:: get_byte_ranges` and short circuit to return empty bytes when this is true. Without this, it will still try to read byte ranges and try to index on an empty vec. Reproducible by the below when the data in S3 is an empty table, see issue for more details. ``` df = daft.read_parquet(<insert s3 path) df.where(~daft.col(<some bool col>)).count_rows() ``` ## Related Issues https://github.com/Eventual-Inc/Daft/issues/6601

main

5 hours ago

skip getting bytes when range start equals end

gweaverbiodev:gw/async-read-range-check

23 hours ago

[update] pre-commit

YannByron:paimon-enhancements

1 day ago

[update] fix ci

YannByron:paimon-enhancements

1 day ago

[update] fix ci

YannByron:paimon-enhancements

1 day ago

Latest Branches

-1%

chore(deps): bump compatible dependencies from dependabot#6603

4 hours ago

5d5f0ef

desmond/fix-dependabot-6598

0%

fix(scan): skip getting bytes when range start equals end in daft async reader#6602

23 hours ago

ff406e1

gweaverbiodev:gw/async-read-range-check

+1%

feat(paimon): enhance Paimon integration with Daft#6553

1 day ago

a7aa571

YannByron:paimon-enhancements

© 2026 CodSpeed Technology

Home Terms Privacy Docs