Apache Arrow 0.17.1 (18 May 2020)
This is a patch release fixing bugs and regressions listed in the changelog below.
Download
- Source Artifacts
- Binary Artifacts
- Git tag
Contributors
This release includes 27 commits from 11 distinct contributors.
$ git shortlog -sn apache-arrow-0.17.0..apache-arrow-0.17.1
9 Krisztián Szűcs
4 Sutou Kouhei
3 Neal Richardson
3 Wes McKinney
2 Antoine Pitrou
1 Eric Erhardt
1 Joris Van den Bossche
1 Micah Kornfield
1 Paul Taylor
1 Yibo Cai
1 Yuan Zhou
Changelog
New Features and Improvements
- ARROW-8501 - [Packaging][RPM] Upgrade devtoolset to 8 on CentOS 6
- ARROW-8549 - [R] Assorted post-0.17 release cleanups
- ARROW-8699 - [R] Fix automatic r_to_py conversion
- ARROW-8758 - [R] Updates for compatibility with dplyr 1.0
- ARROW-8786 - [Packaging][rpm] Use bundled zstd in the CentOS 8 build
Bug Fixes
- ARROW-8503 - [Packaging][deb] Can’t build apache-arrow-archive-keyring for RC
- ARROW-8505 - [Release][C#] “sourcelink test” is failed by Apache.Arrow.AssemblyInfo.cs
- ARROW-8584 - [Packaging][C++] Protobuf link error in deb builds
- ARROW-8608 - [C++] Update vendored mpark/variant.h to latest to fix NVCC compilation issues
- ARROW-8609 - [C++] ORC JNI bridge crashed on null arrow buffer
- ARROW-8641 - [Python] Regression in feather: no longer supports permutation in column selection
- ARROW-8657 - [Python][C++][Parquet] Forward compatibility issue from 0.16 to 0.17 when using version=’2.0’
- ARROW-8684 - [Python] “SystemError: Bad call flags in _PyMethodDef_RawFastCallDict” in Python 3.7.7 on macOS when using pyarrow wheel
- ARROW-8694 - [Python][Parquet] parquet.read_schema() fails when loading wide table created from Pandas DataFrame
- ARROW-8704 - [C++] Fix Parquet crash on invalid input (OSS-Fuzz)
- ARROW-8706 - [C++][Parquet] Tracking JIRA for PARQUET-1857 (unencrypted INT16_MAX Parquet row group limit)
- ARROW-8728 - [C++] Bitmap operation may cause buffer overflow
- ARROW-8741 - [Python][Packaging] Keep VS2015 with for the windows wheels
- ARROW-8750 - [Python] pyarrow.feather.write_feather does not default to lz4 compression if it’s available
- PARQUET-1857 - [C++][Parquet] ParquetFileReader unable to read files with more than 32767 row groups