33 Pull requests merged by 12 people

• Improve high-multiplicity joins benchmark

  #19287 merged Jul 9, 2025

• Enable chunked reading of PQ sources with >2B rows

  #19245 merged Jul 9, 2025

• Clean up cudf._lib.strings_udf.pyx

  #19335 merged Jul 9, 2025

• Add support for pandas-2.3.1

  #19334 merged Jul 9, 2025

• Add single-file streaming Sink support

  #19317 merged Jul 9, 2025

• Refactor hash join with multiset

  #18021 merged Jul 9, 2025

• Fix null mask assignment in aggregators and cleanup with C++20

  #19302 merged Jul 9, 2025

• Allow setting StreamingExecutor.target_partition_size with an environment variable

  #19316 merged Jul 9, 2025

• Optimize object listing in pandas-tests diff CI

  #19328 merged Jul 9, 2025

• Raised MixedTypeErrors for condition that lead to mixed types

  #19232 merged Jul 9, 2025

• Serialize ConfigOptions in pdsh benchmark output

  #19272 merged Jul 9, 2025

• Support Expr.str.find & Expr.str.join for non string data in cudf_polars

  #19275 merged Jul 9, 2025

• Fix job filters for pandas-tests

  #19322 merged Jul 9, 2025

• [pre-commit.ci] pre-commit autoupdate

  #19301 merged Jul 9, 2025

• Autodoc DateOffset

  #19297 merged Jul 8, 2025

• Add support for StructFunction expressions in cudf_polars

  #19052 merged Jul 8, 2025

• Always represent datetime aware data as UTC in strftime

  #19304 merged Jul 8, 2025

• Support Expr.str.extract/extract_groups in cudf_polars

  #19271 merged Jul 8, 2025

• Remove pytest pin

  #19127 merged Jul 8, 2025

• Move ast expression function definitions to .cpp files

  #19250 merged Jul 8, 2025

• Update cudf.pandas test skips for pandas==2.3.1

  #19313 merged Jul 8, 2025

• Fix cudf::column_device_view::element() doxygen

  #19296 merged Jul 8, 2025

• Leverage new pylibcudf grouped_range_rolling_window for cuDF classic rolling(window: timedelta)

  #19230 merged Jul 8, 2025

• Fixed type annotation for 'state' in make_recursive

  #19294 merged Jul 7, 2025

• Implement row group pruning with dictionaries in experimental PQ reader

  #18836 merged Jul 7, 2025

• Add some basic streaming engine documentation

  #19088 merged Jul 7, 2025

• Raise NotImplementedError for LazyFrame.profile with the streaming exeuctor

  #19257 merged Jul 7, 2025

• Move shuffle method defaulting to config options creation

  #19274 merged Jul 7, 2025

• Do not pass cupy objects objects to numba kernels directly

  #19283 merged Jul 3, 2025

• Support Expr.str.json_path_match/len_bytes/len_chars in cudf_polars

  #19277 merged Jul 3, 2025

• Document aggregations for cudf::reduce in doxygen

  #19264 merged Jul 3, 2025

• Rename "cardinality_factor" configuration to "unique_fraction"

  #19273 merged Jul 3, 2025

• Move get_mask_offset_word utility to null_mask.cuh

  #19226 merged Jul 3, 2025

25 Pull requests opened by 10 people

• Latest cuco commit 541bdc0

  #19288 opened Jul 3, 2025

• Add cudf::strings::split_part API

  #19289 opened Jul 3, 2025

• Support Expr.str.splitn/split_exact in cudf_polars

  #19290 opened Jul 3, 2025

• Deprecate cudf::round for float types

  #19298 opened Jul 7, 2025

• Refactor Source Type for UDFs

  #19300 opened Jul 7, 2025

• Add new cudf::top_k API

  #19303 opened Jul 7, 2025

• Move the Parquet `reader_impl` class declaration out of the `parquet::detail::reader`

  #19305 opened Jul 7, 2025

• Support Expr.str.json_decode in cudf_polars

  #19307 opened Jul 8, 2025

• 🚧 Materialize tables in the experimental Parquet reader

  #19308 opened Jul 8, 2025

• Support the JNI build with nvCOMP 5.0

  #19309 opened Jul 8, 2025

• Fix a use-after-free issue in TDigest aggregation code.

  #19311 opened Jul 8, 2025

• Support null_count expression

  #19314 opened Jul 8, 2025

• Remove unnecessary compute for integer windows

  #19315 opened Jul 8, 2025

• Support fill_null with fill strategy in cudf-polars

  #19318 opened Jul 8, 2025

• [DO NOT MERGE] Implement JIT Conditional Join

  #19319 opened Jul 8, 2025

• Fix compile warning in interop_stringview.cpp

  #19320 opened Jul 8, 2025

• [TEST-ONLY] test cuda/std/span in cuco

  #19321 opened Jul 8, 2025

• [TEST-ONLY] test just including span

  #19325 opened Jul 8, 2025

• Add cudf::strings::find_instance API

  #19326 opened Jul 8, 2025

• Fix bit shift overflow in segmented_offset_bitmask_binop utility

  #19329 opened Jul 9, 2025

• Re-enable std/var reductions for libcudf debug builds

  #19331 opened Jul 9, 2025

• Allow comparison binop to datetime.date

  #19333 opened Jul 9, 2025

• Fix Union-Slice bug

  #19336 opened Jul 9, 2025

• Update rapids_config files to support new branching strategy

  #19337 opened Jul 9, 2025

• Support rank expression in cudf-polars

  #19340 opened Jul 10, 2025

20 Issues closed by 3 people

• [BUG] Parquet reader fails to read a chunk for parquet file with more than 2^32 rows

  #19238 closed Jul 9, 2025

• [BUG] assignment through loc[] breaks DataFrame

  #12505 closed Jul 9, 2025

• [FEA] Support `polars.Expr.str.find` in `cudf-polars`

  #18993 closed Jul 9, 2025

• [FEA] Support `polars.Expr.str.join` in `cudf-polars`

  #18996 closed Jul 9, 2025

• DOCS: document `DateOffset`

  #19292 closed Jul 8, 2025

• [FEA] Support struct data types in Polars

  #16725 closed Jul 8, 2025

• [BUG] `dt.strftime` ignores time zone

  #19295 closed Jul 8, 2025

• [FEA] Support `polars.Expr.str.extract_groups` in `cudf-polars`

  #18991 closed Jul 8, 2025

• [FEA] Support `polars.Expr.str.extract` in `cudf-polars`

  #18989 closed Jul 8, 2025

• [BUG] Element access in column_device_view leads to copy

  #19291 closed Jul 8, 2025

• [BUG] `RollingGroupBy` objects cannot be pickled

  #14845 closed Jul 8, 2025

• [Story] Rewrite cudf.pandas rolling functionality using the new C++ rolling functionality

  #18709 closed Jul 8, 2025

• [FEA] `reprog_device::create` and `regex_device_builder::create_prog_device` should receive an `rmm::device_async_resource_ref` argument

  #19286 closed Jul 7, 2025

• [FEA] Add a custom decoder to convert Parquet dictionary pages into `cuco::static_set` objects

  #18046 closed Jul 7, 2025

• [BUG]: `TypeError`: in `LazyFrame.profile()` with streaming executor

  #19253 closed Jul 7, 2025

• [QST]cuda12.8

  #19260 closed Jul 7, 2025

• RUN TPCH 10K on DGXA100/DGXH100/DGXB100

  #18812 closed Jul 3, 2025

• [FEA] Support `polars.Expr.str.len_chars` in `cudf-polars`

  #19000 closed Jul 3, 2025

• [FEA] Support `polars.Expr.str.len_bytes` in `cudf-polars`

  #18999 closed Jul 3, 2025

• [FEA] Support `polars.Expr.str.json_path_match` in `cudf-polars`

  #18998 closed Jul 3, 2025

12 Issues opened by 6 people

• [FEA] Add a `pylibcudf.Column.as_struct(columns: Iterable[Column])` to create a struct column

  #19339 opened Jul 9, 2025

• [FEA] Support string column to struct column (and vice versa) conversions in `read_json`/`write_json`

  #19338 opened Jul 9, 2025

• [BUG] Incorrect result from `tail(1)` after `concat` with streaming GPU executor

  #19332 opened Jul 9, 2025

• [FEA] Allow setting all cudf-polars configuration options with environment variables

  #19330 opened Jul 9, 2025

• [FEA] support comparison between DatetimeColumn and native python date object

  #19327 opened Jul 9, 2025

• [FEA] Allow controlling rounding behavior for `quantile(interp="nearest")` to match Polars

  #19324 opened Jul 8, 2025

• [BUG] Difference in `groupby().all()` with `cudf.pandas` and string dtype

  #19312 opened Jul 8, 2025

• [FEA] Single-file ``Sink`` support for streaming executor

  #19310 opened Jul 8, 2025

• [FEA] Create post op CPU callback for cudf_polars output types that are not supported by libcudf(?)

  #19306 opened Jul 8, 2025

• [FEA] `contiguous_split` should accept span of split points, rather than vector

  #19293 opened Jul 7, 2025

• [FEA] Support `Expr.name.map_fields` in cudf_polars

  #19285 opened Jul 3, 2025

• [FEA] Support `Expr.struct.with_fields` in cudf_polars

  #19284 opened Jul 3, 2025

34 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Implement UDF Filters

  #19070 commented on Jul 9, 2025 • 40 new comments

• Introduce classes for collecting source statistics

  #19276 commented on Jul 9, 2025 • 23 new comments

• Change default cudf-polars executor to "streaming"

  #19263 commented on Jul 9, 2025 • 6 new comments

• Added Concurrent Polynomial Transform Benchmark

  #19199 commented on Jul 8, 2025 • 4 new comments

• Manage strings with NRT

  #18453 commented on Jul 7, 2025 • 4 new comments

• Support output_dtype in cudf::reduce for nunique aggregation

  #19265 commented on Jul 9, 2025 • 3 new comments

• Support ternary expression inside groupby context

  #19242 commented on Jul 8, 2025 • 2 new comments

• Experimental API to read a parquet table, build a custom index column, and apply roaring bitmap deletion vector

  #19237 commented on Jul 7, 2025 • 2 new comments

• Add SUM_ANSI aggregation for INT64

  #19282 commented on Jul 3, 2025 • 1 new comment

• Support Expr.str.strip_prefix/suffix in cudf_polars

  #19278 commented on Jul 8, 2025 • 1 new comment

• Require `numba-cuda>=0.16.0`

  #19213 commented on Jul 9, 2025 • 1 new comment

• [BUG] Assertion error when running JOIN_NVBENCH with debug build

  #19255 commented on Jul 3, 2025 • 0 new comments

• Add data types axis to joins benchmarks

  #19281 commented on Jul 7, 2025 • 0 new comments

• [FEA] overflow detection for SUM aggregations

  #19243 commented on Jul 3, 2025 • 0 new comments

• [FEA] Support `polars.Expr.str.split_exact` in `cudf-polars`

  #19027 commented on Jul 3, 2025 • 0 new comments

• Fix includes for segmented-reduce source files

  #19266 commented on Jul 8, 2025 • 0 new comments

• [FEA] Support `polars.Expr.str.zfill` in `cudf-polars`

  #19035 commented on Jul 3, 2025 • 0 new comments

• [FEA] Support `polars.Expr.str.pad_start` in `cudf-polars`

  #19024 commented on Jul 3, 2025 • 0 new comments

• [FEA] Support `polars.Expr.str.pad_end` in `cudf-polars`

  #19023 commented on Jul 3, 2025 • 0 new comments

• [FEA] Support `polars.Expr.str.strip_prefix` in `cudf-polars`

  #19029 commented on Jul 3, 2025 • 0 new comments

• Remove cudautils.py

  #19233 commented on Jul 9, 2025 • 0 new comments

• [FEA] Support `polars.Expr.str.splitn` in `cudf-polars`

  #19028 commented on Jul 3, 2025 • 0 new comments

• [FEA] Support `polars.Expr.str.strip_suffix` in `cudf-polars`

  #19030 commented on Jul 4, 2025 • 0 new comments

• Use GCC 14 in conda builds.

  #19192 commented on Jul 7, 2025 • 0 new comments

• Use KvikIO's implementation of file-backed memory mapping

  #19164 commented on Jul 7, 2025 • 0 new comments

• [FEA] Support NaN-aware operators in libcudf that conform to ETL engine expectations

  #18930 commented on Jul 7, 2025 • 0 new comments

• Add multi-column support for primitive row operator dispatch

  #18940 commented on Jul 3, 2025 • 0 new comments

• [BUG] Uninitialized read detected in parquet chunked reader

  #19247 commented on Jul 7, 2025 • 0 new comments

• [FEA] Support `polars.Expr.str.json_decode` in `cudf-polars`

  #18997 commented on Jul 9, 2025 • 0 new comments

• [BUG] Rolling window aggregations are very slow with large windows

  #15119 commented on Jul 8, 2025 • 0 new comments

• [FEA] Support `polars.Expr.str.to_decimal` in `cudf-polars`

  #19032 commented on Jul 8, 2025 • 0 new comments

• [FEA] Support `polars.Expr.str.decode` in `cudf-polars`

  #18986 commented on Jul 8, 2025 • 0 new comments

• [FEA] Support `polars.Expr.str.to_time` in `cudf-polars`

  #19033 commented on Jul 8, 2025 • 0 new comments

• [FEA] Add a TopK API to cudf

  #19096 commented on Jul 8, 2025 • 0 new comments