v23.04.01

🚨 Breaking Changes

• Pin dask and distributed for release (#13070) @galipremsagar
• Declare a different name for nan_equality.UNEQUAL to prevent Cython warnings. (#12947) @bdice
• Update minimum pandas and numpy pinnings (#12887) @galipremsagar
• Deprecate names & dtype in Index.copy (#12825) @galipremsagar
• Deprecate Index.is_* methods (#12820) @galipremsagar
• Deprecate datetime_is_numeric from describe (#12818) @galipremsagar
• Deprecate na_sentinel in factorize (#12817) @galipremsagar
• Make string methods return a Series with a useful Index (#12814) @shwina
• Produce useful guidance on overflow error in to_csv (#12705) @wence-
• Move strings_udf code into cuDF (#12669) @brandon-b-miller
• Remove cudf::strings::repeat_strings_output_sizes and optional parameter from cudf::strings::repeat_strings (#12609) @davidwendt
• Replace message parsing with throwing more specific exceptions (#12426) @vyasr

🐛 Bug Fixes

• Pin curand version (#13127) @vyasr
• Fix memcheck script to execute only _TEST files found in bin/gtests/libcudf (#13006) @davidwendt
• Fix DataFrame constructor to broadcast scalar inputs properly (#12997) @galipremsagar
• Drop force_nullable_schema from chunked parquet writer (#12996) @galipremsagar
• Fix gtest column utility comparator diff reporting (#12995) @davidwendt
• Handle index names while performing groupby (#12992) @galipremsagar
• Fix __setitem__ on string columns when the scalar value ends in a null byte (#12991) @wence-
• Fix sort_values when column is all empty strings (#12988) @eriknw
• Remove unused variable and fix memory issue in ORC writer (#12984) @ttnghia
• Pre-emptive fix for upstream dask.dataframe.read_parquet changes (#12983) @rjzamora
• Remove MANIFEST.in use auto-generated one for sdists and package_data for wheels (#12960) @vyasr
• Update to use rapids-export(COMPONENTS) feature. (#12959) @robertmaynard
• cudftestutil supports static gtest dependencies (#12957) @robertmaynard
• Include gtest in build environment. (#12956) @vyasr
• Correctly handle scalar indices in Index.__getitem__ (#12955) @wence-
• Avoid building cython twice (#12945) @galipremsagar
• Fix set index error for Series rolling window operations (#12942) @galipremsagar
• Fix calculation of null counts for Parquet statistics (#12938) @etseidl
• Preserve integer dtype of hive-partitioned column containing nulls (#12930) @rjzamora
• Use get_current_device_resource for intermediate allocations in COLLECT_LIST window code (#12927) @karthikeyann
• Mark dlpack tensor deleter as noexcept to match PyCapsule_Destructor signature. (#12921) @bdice
• Fix conda recipe post-link.sh typo (#12916) @pentschev
• min_rows and num_rows are swapped in ComputePageSizes declaration in Parquet reader (#12886) @etseidl
• Expect cupy to now support bool arrays for dlpack. (#12883) @vyasr
• Use python -m pytest for nightly wheel tests (#12871) @bdice
• Parquet writer column_size() should return a size_t (#12870) @etseidl
• Fix cudf::hash_partition kernel launch error with decimal128 types (#12863) @davidwendt
• Fix an issue with parquet chunked reader undercounting string lengths. (#12859) @nvdbaranec
• Remove tokenizers pre-install pinning. (#12854) @vyasr
• Fix parquet RangeIndex bug (#12838) @rjzamora
• Remove KAFKA_HOST_TEST from compute-sanitizer check (#12831) @davidwendt
• Make string methods return a Series with a useful Index (#12814) @shwina
• Tell cudf_kafka to use header-only fmt (#12796) @vyasr
• Add GroupBy.dtypes (#12783) @galipremsagar
• Fix a leak in a test and clarify some test names (#12781) @revans2
• Fix bug in all-null list due to join_list_elements special handling (#12767) @karthikeyann
• Add try/except for expected null-schema error in read_parquet (#12756) @rjzamora
• Throw an exception if an unsupported page encoding is detected in Parquet reader (#12754) @etseidl
• Fix a bug with num_keys in _scatter_by_slice (#12749) @thomcom
• Bump pinned rapids wheel deps to 23.4 (#12735) @sevagh
• Rework logic in cudf::strings::split_record to improve performance (#12729) @davidwendt
• Add always_nullable flag to Dremel encoding (#12727) @divyegala
• Fix memcheck read error in compound segmented reduce (#12722) @davidwendt
• Fix faulty conditional logic in JIT GroupBy.apply (#12706) @brandon-b-miller
• Produce useful guidance on overflow error in to_csv (#12705) @wence-
• Handle parquet list data corner case (#12698) @nvdbaranec
• Fix missing trailing comma in json writer (#12688) @karthikeyann
• Remove child fom newCudaAsyncMemoryResource (#12681) @abellina
• Handle bool types in round API (#12670) @galipremsagar
• Ensure all of device bitmask is initialized in from_arrow (#12668) @wence-
• Fix from_arrow to load a sliced arrow table (#12665) @galipremsagar
• Fix dask-cudf read_parquet bug for multi-file aggregation (#12663) @rjzamora
• Fix AllocateLikeTest gtests reading uninitialized null-mask (#12643) @davidwendt
• Fix find_common_dtype and values to handle complex dtypes (#12537) @galipremsagar
• Fix fetching of MultiIndex values when a label is passed (#12521) @galipremsagar
• Fix Series comparison vs scalars (#12519) @brandon-b-miller
• Allow casting from UDFString back to StringView to call methods in strings_udf (#12363) @brandon-b-miller

📖 Documentation

• Fix GroupBy.apply doc examples rendering (#12994) @brandon-b-miller
• add sphinx building and s3 uploading for dask-cudf docs (#12982) @quasiben
• Add developer documentation forbidding default parameters in detail APIs (#12978) @vyasr
• Add README symlink for dask-cudf. (#12946) @bdice
• Remove return type from @return doxygen tags (#12908) @davidwendt
• Fix docs build to be pydata-sphinx-theme=0.13.0 compatible (#12874) @galipremsagar
• Add skeleton API and prose documentation for dask-cudf (#12725) @wence-
• Enable doctests for GroupBy methods (#12658) @brandon-b-miller
• Add comment about CUB patch for SegmentedSortInt.Bool gtest (#12611) @davidwendt

🚀 New Features

• Add JNI method for strings::replace multi variety (#12979) @NVnavkumar
• Add nunique aggregation support for cudf::segmented_reduce (#12972) @davidwendt
• Refactor orc chunked writer (#12949) @ttnghia
• Make Parquet writer nullable option application to single table writes (#12933) @vuule
• Refactor io::orc::ProtobufWriter (#12877) @ttnghia
• Make timezone table independent from ORC (#12805) @vuule
• Cache JIT GroupBy.apply functions (#12802) @brandon-b-miller
• Implement initial support for avro logical types (#6482) (#12788) @tpn
• Update tests/column_utilities to use experimental::equality row comparator (#12777) @divyegala
• Update distinct/unique_count to experimental::row hasher/comparator (#12776) @divyegala
• Update hash_partition to use experimental::row::row_hasher (#12761) @divyegala
• Update is_sorted to use experimental::row::lexicographic (#12752) @divyegala
• Update default data source in cuio reader benchmarks (#12740) @PointKernel
• Reenable stream identification library in CI (#12714) @vyasr
• Add regex_program strings splitting java APIs and tests (#12713) @cindyyuanjiang
• Add regex_program strings replacing java APIs and tests (#12701) @cindyyuanjiang
• Add regex_program strings extract java APIs and tests (#12699) @cindyyuanjiang
• Variable fragment sizes for Parquet writer (#12685) @etseidl
• Add segmented reduction support for fixed-point types (#12680) @davidwendt
• Move strings_udf code into cuDF (#12669) @brandon-b-miller
• Add regex_program searching APIs and related java classes (#12666) @cindyyuanjiang
• Add logging to libcudf (#12637) @vuule
• Add compound aggregations to cudf::segmented_reduce (#12573) @davidwendt
• Convert rank to use to experimental row comparators (#12481) @divyegala
• Use rapids-cmake parallel testing feature (#12451) @robertmaynard
• Enable detection of undesired stream usage (#12089) @vyasr

🛠️ Improvements

• Pin dask and distributed for release (#13070) @galipremsagar
• Pin cupy in wheel tests to supported versions (#13041) @vyasr
• Pin numba version (#13001) @vyasr
• Rework gtests SequenceTest to remove using namepace cudf (#12985) @davidwendt
• Stop setting package version attribute in wheels (#12977) @vyasr
• Move detail reduction functions to cudf::reduction::detail namespace (#12971) @davidwendt
• Remove default detail mrs: part7 (#12970) @vyasr
• Remove default detail mrs: part6 (#12969) @vyasr
• Remove default detail mrs: part5 (#12968) @vyasr
• Remove default detail mrs: part4 (#12967) @vyasr
• Remove default detail mrs: part3 (#12966) @vyasr
• Remove default detail mrs: part2 (#12965) @vyasr
• Remove default detail mrs: part1 (#12964) @vyasr
• Add force_nullable_schema parameter to Parquet writer. (#12952) @galipremsagar
• Declare a different name for nan_equality.UNEQUAL to prevent Cython warnings. (#12947) @bdice
• Remove remaining default stream parameters (#12943) @vyasr
• Fix cudf::segmented_reduce gtest for ANY aggregation (#12940) @davidwendt
• Implement groupby.head and groupby.tail (#12939) @wence-
• Fix libcudf gtests to pass null-count=0 for empty validity masks (#12923) @davidwendt
• Migrate parquet encoding to use experimental row operators (#12918) @PointKernel
• Fix benchmarks coded in namespace cudf and using namespace cudf (#12915) @karthikeyann
• Fix io/text gtests coded in namespace cudf::test (#12914) @karthikeyann
• Pass SCCACHE_S3_USE_SSL to conda builds (#12910) @ajschmidt8
• Fix FST, JSON gtests & benchmarks coded in namespace cudf::test (#12907) @karthikeyann
• Generate pyproject dependencies using dfg (#12906) @vyasr
• Update libcudf counting functions to specify cudf::size_type (#12904) @davidwendt
• Fix moto env vars & pass AWS_SESSION_TOKEN to conda builds (#12902) @ajschmidt8
• Rewrite CSV writer benchmark with nvbench (#12901) @PointKernel
• Rework some code logic to reduce iterator and comparator inlining to improve compile time (#12900) @davidwendt
• Deprecate line_terminator in favor of lineterminator in to_csv (#12896) @wence-
• Add stream and mr parameters for structs::detail::flatten_nested_columns (#12892) @ttnghia
• Deprecate libcudf regex APIs accepting pattern strings directly (#12891) @davidwendt
• Remove default parameters from detail headers in include (#12888) @vyasr
• Update minimum pandas and numpy pinnings (#12887) @galipremsagar
• Implement groupby.sample (#12882) @wence-
• Update JNI build ENV default to gcc 11 (#12881) @pxLi
• Change return type of cudf::structs::detail::flatten_nested_columns to smart pointer (#12878) @ttnghia
• Fix passing seed parameter to MurmurHash3_32 in cudf::hash() function (#12875) @davidwendt
• Remove manual artifact upload step in CI (#12869) @ajschmidt8
• Update to GCC 11 (#12868) @bdice
• Fix null hive-partition behavior in dask-cudf parquet (#12866) @rjzamora
• Update to protobuf>=4.21.6,<4.22. (#12864) @bdice
• Update RMM allocators (#12861) @pentschev
• Improve performance for replace-multi for long strings (#12858) @davidwendt
• Drop Python 3.7 handling for pickle protocol 4 (#12857) @jakirkham
• Migrate as much as possible to pyproject.toml (#12850) @vyasr
• Enable nbqa pre-commit hooks for isort and black. (#12848) @bdice
• Setting a threshold for KvikIO IO (#12841) @madsbk
• Update datasets download URL (#12840) @jjacobelli
• Make docs builds less verbose (#12836) @AyodeAwe
• Consolidate linter configs into pyproject.toml (#12834) @vyasr
• Deprecate names & dtype in Index.copy (#12825) @galipremsagar
• Deprecate inplace parameters in categorical methods (#12824) @galipremsagar
• Add optional text file support to ninja-log utility (#12823) @davidwendt
• Deprecate Index.is_* methods (#12820) @galipremsagar
• Add dfg as a pre-commit hook (#12819) @vyasr
• Deprecate datetime_is_numeric from describe (#12818) @galipremsagar
• Deprecate na_sentinel in factorize (#12817) @galipremsagar
• Shuffling read into a sub function in parquet read (#12809) @hyperbolic2346
• Fixing parquet coalescing of reads (#12808) @hyperbolic2346
• CI: Remove specification of manual stage for check_style.sh script. (#12803) @csadorf
• Add compute-sanitizer github workflow action to nightly tests (#12800) @davidwendt
• Enable groupby std and variance aggregation types in libcudf Debug build (#12799) @davidwendt
• Expose seed argument to hash_values (#12795) @ayushdg
• Fix groupby gtests coded in namespace cudf::test (#12784) @davidwendt
• Improve performance for cudf::strings::count_characters for long strings (#12779) @davidwendt
• Deallocate encoded data in ORC writer immediately after compression (#12770) @vuule
• Stop force pulling fmt in nvbench. (#12768) @vyasr
• Remove now redundant cuda initialization (#12758) @vyasr
• Adds JSON reader, writer io benchmark (#12753) @karthikeyann
• Use test paths relative to package directory. (#12751) @bdice
• Add build metrics report as artifact to cpp-build workflow (#12750) @davidwendt
• Add JNI methods for detecting and purging non-empty nulls from LIST and STRUCT (#12742) @razajafri
• Stop using versioneer to manage versions (#12741) @vyasr
• Reduce error handling verbosity in CI tests scripts (#12738) @AjayThorve
• Reduce the number of test cases in multibyte_split benchmark (#12737) @PointKernel
• Update shared workflow branches (#12733) @ajschmidt8
• JNI switches to nested JSON reader (#12732) @res-life
• Changing cudf::io::source_info to use cudf::host_span<std::byte> in a non-breaking form (#12730) @hyperbolic2346
• Add nvbench environment class for initializing RMM in benchmarks (#12728) @davidwendt
• Split C++ and Python build dependencies into separate lists. (#12724) @bdice
• Add build dependencies to Java tests. (#12723) @bdice
• Allow setting the seed argument for hash partition (#12715) @firestarman
• Remove gpuCI scripts. (#12712) @bdice
• Unpin dask and distributed for development (#12710) @galipremsagar
• partition_by_hash(): use _split() (#12704) @madsbk
• Remove DataFrame.quantiles from docs. (#12684) @bdice
• Fast path for experimental::row::equality (#12676) @divyegala
• Move date to build string in conda recipe (#12661) @ajschmidt8
• Refactor reduction logic for fixed-point types (#12652) @davidwendt
• Pay off some JNI RMM API tech debt (#12632) @revans2
• Merge copy-on-write feature branch into branch-23.04 (#12619) @galipremsagar
• Remove cudf::strings::repeat_strings_output_sizes and optional parameter from cudf::strings::repeat_strings (#12609) @davidwendt
• Pin cuda-nvrtc. (#12606) @bdice
• Remove cudf::test::print calls from libcudf gtests (#12604) @davidwendt
• Init JNI version 23.04.0-SNAPSHOT (#12599) @pxLi
• Add performance benchmarks to user facing docs (#12595) @galipremsagar
• Add docs build job (#12592) @AyodeAwe
• Replace message parsing with throwing more specific exceptions (#12426) @vyasr
• Support conversion to/from cudf in dask.dataframe.core.to_backend (#12380) @rjzamora