Releases: G-Research/ParquetSharp
ParquetSharp 23.0.0.2
This is a bug fix release that statically links libatomic on Linux, which was introduced as an extra dependency of the ParquetSharp native library in version 23.0.0.
What's Changed
- Remove libatomic.so from runtime dependencies by @marcin-krystianc in #643
Full Changelog: 23.0.0.1...23.0.0.2
ParquetSharp 23.0.0.1
This is a bug-fix release that fixes a potential stack overflow when reading very wide decimal typed columns. It's recommended to upgrade to this version if you read untrusted Parquet files.
What's Changed
- Fix documentation for
WriterPropertiesBuilder.MaxRowGroupLengthby @mobiusklein in #641 - Add a limit on stackalloc size when reading Decimal columns by @adamreeve in #642
New Contributors
- @mobiusklein made their first contribution in #641
Full Changelog: 23.0.0...23.0.0.1
ParquetSharp 23.0.0
This version of ParquetSharp upgrades the underlying C++ Parquet library to 23.0.0.
What's Changed
- Add
ThriftStringSizeLimittoReaderPropertiesby @haianhng31 in #564 - Add
ThriftContainerSizeLimitandFooterReadSizetoReaderPropertiesby @haianhng31 in #567 - Add support for Windows Arm64 by @jgiannuzzi in #576
- Add
BinaryTypetoArrowReaderPropertiesby @haianhng31 in #569 - Add
ListTypeandArrowExtensionsEnabledtoArrowReaderPropertiesby @haianhng31 in #582 - Add
StoreDecimalAsIntegertoWriterPropertiesby @haianhng31 in #572 - Add
SetMaxStatisticsSizetoWriterPropertiesBuilderby @haianhng31 in #581 - Add
DataPageVersionandSizeStatisticsLeveltoWriterPropertiesby @haianhng31 in #585 - Add
CacheOptionstoArrowReaderPropertiesby @haianhng31 in #589 - Add field_id to
Columnby @acolombi in #625 - Upgrade C++ arrow to 23.0.0 by @marcin-krystianc in #626
- Add Memory-Optimized Reading Benchmarks documentation by @larrytamnjong in #611
New Contributors
- @haianhng31 made their first contribution in #565
- @iv-tunate made their first contribution in #597
- @acolombi made their first contribution in #625
- @larrytamnjong made their first contribution in #611
Full Changelog: 21.0.0...23.0.0
ParquetSharp 23.0.0-beta1
This version of ParquetSharp upgrades the underlying C++ Parquet library to 23.0.0.
What's Changed
- Add
ThriftStringSizeLimittoReaderPropertiesby @haianhng31 in #564 - Add
ThriftContainerSizeLimitandFooterReadSizetoReaderPropertiesby @haianhng31 in #567 - Add support for Windows Arm64 by @jgiannuzzi in #576
- Add
BinaryTypetoArrowReaderPropertiesby @haianhng31 in #569 - Add
ListTypeandArrowExtensionsEnabledtoArrowReaderPropertiesby @haianhng31 in #582 - Add
StoreDecimalAsIntegertoWriterPropertiesby @haianhng31 in #572 - Add
SetMaxStatisticsSizetoWriterPropertiesBuilderby @haianhng31 in #581 - Add
DataPageVersionandSizeStatisticsLeveltoWriterPropertiesby @haianhng31 in #585 - Add
CacheOptionstoArrowReaderPropertiesby @haianhng31 in #589 - Add field_id to
Columnby @acolombi in #625 - Upgrade C++ arrow to 23.0.0 by @marcin-krystianc in #626
- Add Memory-Optimized Reading Benchmarks documentation by @larrytamnjong in #611
New Contributors
- @haianhng31 made their first contribution in #565
- @iv-tunate made their first contribution in #597
- @acolombi made their first contribution in #625
- @larrytamnjong made their first contribution in #611
Full Changelog: 21.0.0...23.0.0-beta1
ParquetSharp 21.0.0
What's Changed
- Add sorting columns by @iksarfo in #536
- Upgrade C++ Arrow to 21.0.0 by @adamreeve in #548
- Allow specifying the native allocator in code by @adamreeve in #557
New Contributors
Full Changelog: 20.0.0...21.0.0
ParquetSharp 21.0.0-beta1
What's Changed
- Add sorting columns by @iksarfo in #536
- Upgrade C++ Arrow to 21.0.0 by @adamreeve in #548
- Allow specifying the native allocator in code by @adamreeve in #557
New Contributors
Full Changelog: 20.0.0...21.0.0-beta1
ParquetSharp 20.0.0
This version of ParquetSharp upgrades the underlying Arrow C++ Parquet library to version 20.0.0. ParquetSharp also now builds Arrow with support for the mimalloc and jemalloc native memory allocators, and will use mimalloc by default. You can switch back to the old system allocator by setting the ARROW_DEFAULT_MEMORY_POOL environment variable to system.
Note that if writing large Parquet files that use array typed columns, there is a known bug that can cause page sizes to grow large. This can be worked around by disabling page index writing.
What's Changed
- Support writing custom types that can represent nulls by @adamreeve in #526
- Support writing required byte arrays by @adamreeve in #531
- Upgrade Arrow C++ to 20.0.0 by @adamreeve in #528
- Don't require passing a logical type override for Guid typed columns by @adamreeve in #535
- Use ArrayPool for Reader and Writer Buffers by @johnthcall in #537
- Build Arrow with support for mimalloc and jemalloc by @adamreeve in #540
Full Changelog: 19.0.1...20.0.0
ParquetSharp 20.0.0-beta2
What's Changed
- Use ArrayPool for Reader and Writer Buffers by @johnthcall in #537
- Build Arrow with support for mimalloc and jemalloc by @adamreeve in #540
Full Changelog: 20.0.0-beta1...20.0.0-beta2
ParquetSharp 20.0.0-beta1
This version of ParquetSharp upgrades the underlying Arrow C++ Parquet library to version 20.0.0.
What's Changed
- Support writing custom types that can represent nulls by @adamreeve in #526
- Support writing required byte arrays by @adamreeve in #531
- Upgrade Arrow C++ to 20.0.0 by @adamreeve in #528
- Don't require passing a logical type override for Guid typed columns by @adamreeve in #535
Full Changelog: 19.0.1...20.0.0-beta1
ParquetSharp 19.0.1
This version of ParquetSharp upgrades the underlying Arrow C++ Parquet library to version 19.0.1.
What's Changed
- Add support for setting field_id on nodes by @CurtHagenlocher in #509
- Enable Control Flow Guard for windows builds by @johnthcall in #516
- Upgrade C++ Arrow to 19.0.1 by @marcin-krystianc in #520
New Contributors
- @CurtHagenlocher made their first contribution in #509
- @johnthcall made their first contribution in #516
Full Changelog: 18.1.0...19.0.1