What's Changed

• Unpin dask/distributed for development by @charlesbluca in #564
• Update docs theme by @scharlottej13 in #567
• Make sure scheduler has Dask nightlies in upstream cluster testing by @charlesbluca in #573
• Update gpuCI RAPIDS_VER to 22.08 by @github-actions in #565
• Modify test environment pinnings to cover minimum versions by @charlesbluca in #555
• Don't move jar to local mvn repo by @ksonj in #579
• Add max version constraint for fugue by @charlesbluca in #639
• Add environment file & documentation for GPU tests by @charlesbluca in #633
• Validate UDF metadata by @brandon-b-miller in #641
• Set Dask-sql as the default Fugue Dask engine when installed by @goodwanghan in #640
• Generalize analyze/sample tests to resolve CI failures by @charlesbluca in #668
• Update CodeCov upload step in CI by @charlesbluca in #666
• Bump arrow version to 6.0.0 by @charlesbluca in #674
• Update gpuCI RAPIDS_VER to 22.10 by @github-actions in #665
• Constrain dask pinnings for release by @ayushdg in #690

New Contributors

• @scharlottej13 made their first contribution in #567
• @ksonj made their first contribution in #579

Full Changelog: 2022.6.0...2022.8.0

What's Changed

• Unpin Dask/distributed versions by @charlesbluca in #452
• Add jsonschema to ci testing by @ayushdg in #454
• Switch tests from pd.testing.assert_frame_equal to dd.assert_eq by @charlesbluca in #365
• Set max pin on antlr4-python-runtime by @ayushdg in #456
• Move / minimize number of cudf / dask-cudf imports by @charlesbluca in #480
• Use map_partitions to compute LIMIT / OFFSET by @charlesbluca in #517
• Use dev images for independent cluster testing by @charlesbluca in #518
• Add documentation for FugueSQL integrations by @charlesbluca in #523
• Timestampdiff support by @ayushdg in #495
• Relax jsonschema testing dependency by @charlesbluca in #546
• Update upstream testing workflows by @charlesbluca in #536
• Fix pyarrow / cloudpickle failures in cluster testing by @charlesbluca in #553
• Use bash -l as default entrypoint for all upstream testing jobs by @charlesbluca in #552
• Constrain dask/distributed for release by @charlesbluca in #563

Full Changelog: 2022.4.1...2022.6.0

What's Changed

• Add Java source code to source distribution by @charlesbluca in #451
• Bump httpclient dependency by @charlesbluca in #453

Full Changelog: 2022.4.0...2022.4.1

What's Changed

• Switch github-script action to v3 by @charlesbluca in #379
• Unpin dask/distributed following release by @charlesbluca in #381
• Fix typo by @wence- in #382
• Remove defaults channel from conda envs by @charlesbluca in #384
• Don't persist dataframes before applying offset / limit by @charlesbluca in #387
• Update gpuCI RAPIDS_VER to 22.04 by @github-actions in #374
• Feature/jdbc by @PeterLappo in #351
• Bump gpuCI PYTHON_VER to 3.9 by @charlesbluca in #388
• Stop using defaults channel in dev environments by @charlesbluca in #393
• Use versioneer to compute __version__ by @charlesbluca in #396
• [REVIEW] Modified show.ftl to conditionally expect FROM in parsing logic by @jdye64 in #371
• Fix TIMESTAMP / DATE scalars, add support for DATE column casting by @charlesbluca in #343
• Enable ability for user to pass in a list of CBO rules that should be… by @jdye64 in #389
• Drop support for python 3.7, add testing for python 3.10 by @charlesbluca in #383
• Bump pre-release package versions to be greater than stable releases by @charlesbluca in #405
• Update pytest to generate a client fixture by @ayushdg in #398
• Use build_ext/install_lib subclasses to build external java by @charlesbluca in #406
• Fix use of row UDFs at intermediate query stages by @brandon-b-miller in #409
• [Review] Refactor ConfigContainer to use dask config by @ayushdg in #392
• Provide meta to result of complex _apply_offset by @charlesbluca in #420
• Fix logic for unary join operands like IS NOT NULL by @charlesbluca in #428
• Update docs theme, use sphinx-tabs for CPU/GPU examples by @charlesbluca in #394
• Resolve independent cluster test failures by @charlesbluca in #437
• Only use session-wide client fixture for independent cluster testing by @charlesbluca in #439
• Drop common column from result of cross join, remove from corresponding meta by @charlesbluca in #408
• Add basic predicate-pushdown optimization by @rjzamora in #433
• Add workflow to keep datafusion-sql-planner branch up to date by @charlesbluca in #440
• Update gpuCI RAPIDS_VER to 22.06 by @github-actions in #434
• Bump black style checks to 22.3.0 by @charlesbluca in #443
• Check for ucx-py nightlies when updating gpuCI by @charlesbluca in #441
• Add handling for newer prompt_toolkit versions in cmd tests by @charlesbluca in #447
• Resolve gpuCI workflow failures by @charlesbluca in #446
• Update versions of Java dependencies by @ayushdg in #445
• Update jackson databind version by @ayushdg in #449
• Disable SQL server functionality by @charlesbluca in #448
• Update dask pinnings for release by @charlesbluca in #450

New Contributors

• @wence- made their first contribution in #382
• @PeterLappo made their first contribution in #351
• @rjzamora made their first contribution in #433

Full Changelog: 2022.1.0...2022.4.0

What's Changed

• Disable CodeCov upload in tests on forks by @charlesbluca in #349
• Cost based optimization by @nils-braun in #226
• Add latest dask-ml to upstream testing by @charlesbluca in #354
• Bump gpuCI CUDA_VER to 11.5 by @charlesbluca in #348
• Update Calcite to 1.29.0 and log4j to 2.17.0 to address CVE-2021-44228 by @jdye64 in #347
• Removed uneeded log4j instance that was causing version conflicts and generating slf4j warning messages by @jdye64 in #358
• Added getContext() method to DaskPlanner to ensure that CalciteConfigC… by @jdye64 in #362
• Add os environment option to enable remote jvm debugging by @jdye64 in #363
• Fix issue reporting in scheduled upstream testing by @charlesbluca in #355
• Remove Join Condition Push CBO Rule since it was causing infinite cos… by @jdye64 in #359
• Parse ROWS as tuples in SQL kwargs by @charlesbluca in #338
• Add support for gpu kwarg in Context.sql and explain by @charlesbluca in #368
• Remove max version restriction for Dask/Distributed by @charlesbluca in #369
• Use upstream Dask for complex sorting operations by @charlesbluca in #336
• xfail failing model tests by @charlesbluca in #373
• Add substr tests by @ayushdg in #372
• Fix pandas BaseIndexer import by @charlesbluca in #377
• Bump dask-ml dependency by @charlesbluca in #378
• [REVIEW] Fix unary conditional join operations by @jdye64 in #366
• Pin dask/distributed versions for release by @charlesbluca in #380

Full Changelog: 2021.12.0...2022.1.0

What's Changed

• Update nightly recipe / setup for 2021.11.0 release by @charlesbluca in #308
• Add test build using latest Dask/Distributed by @charlesbluca in #306
• General GHA workflow clean up by @charlesbluca in #313
• Add testing for Python 3.9 by @charlesbluca in #314
• Use Boa for nightly builds by @charlesbluca in #318
• Add handling for cuDF-backed tables in dask-sql-server by @charlesbluca in #312
• Row UDF scalar arguments by @brandon-b-miller in #311
• Update register_func() in context.py by @DaceT in #282
• Bump dask-ml dependency to 2021.11.16 by @charlesbluca in #322
• Add groupby split_out config options to dask-sql by @ayushdg in #286
• Remove null-splitting from _perform_aggregation by @charlesbluca in #273
• Revert "Remove null-splitting from _perform_aggregation" by @charlesbluca in #325
• Resolve failures in nightly package builds by @charlesbluca in #328
• Add workflow to automate gpuCI updates by @charlesbluca in #327
• Update gpuCI RAPIDS_VER to 22.02 by @github-actions in #329
• Installing Dask-SQL w/ RAPIDS by @DaceT in #324
• Remove null-splitting from _perform_aggregation by @charlesbluca in #326
• Generalize table check in _get_tables_from_stack by @charlesbluca in #333
• Add support for GPU table creation in dask / location plugins by @charlesbluca in #251
• Circumvent deep copy of context in PredictModelPlugin by @charlesbluca in #334
• Unrestrict conda-build version used for nightly builds by @charlesbluca in #335
• Update conditions for apply_sort fast codepath by @charlesbluca in #337
• [REVIEW]Add support and tests for cuML and XGBoost by @VibhuJawa in #330
• Ignore case for queries in the parser configuration by @jdye64 in #316
• Ignore .swp files by @charlesbluca in #341
• Added Alter schema and Alter Table by @rajagurunath in #285
• Bump dask dependency to >=2021.11.1,<=2021.11.2 by @charlesbluca in #345

New Contributors

• @DaceT made their first contribution in #282
• @github-actions made their first contribution in #329

Full Changelog: 2021.11.0...2021.12.0

What's Changed

• Use unique names for null/non-null groupby columns by @charlesbluca in #289
• Use string separator in nightly version string by @charlesbluca in #295
• [Review] Update readme and docstrings to indicate GPU support by @ayushdg in #292
• Add DISTRIBUTE BY to dask-sql grammar by @jdye64 in #228
• Use Dask's sort_values for first column sorting in apply_sort by @charlesbluca in #255
• xfail broken dask-ml tests by @charlesbluca in #304
• Bump dask pinning to 2021.10.0 by @charlesbluca in #303
• Prevent JVM Segfault by @jdye64 in #294
• Make meta consistent with results of cross join by @charlesbluca in #300

Full Changelog: 0.4.0...2021.11.0

What's Changed

• More efficient window implementation by @nils-braun in #217
• Support creating tables from cudf dataframes by @ayushdg in #220
• Re-enable the hive tests by @nils-braun in #221
• Reading tables with a dask-cudf DataFrame by @sarahyurick in #224
• Introduces parallel tests to speed up the processing by @nils-braun in #230
• Explicitly install sasl in CI by @charlesbluca in #244
• Add gpuCI support by @charlesbluca in #240
• Add issue templates by @charlesbluca in #247
• Fix test_deprecation_warning in gpuCI by @charlesbluca in #248
• [Review] Add fast path for multi-column sorting by @quasiben in #229
• Add conda dev environments for Python 3.7/3.8, JDK 8/11 by @charlesbluca in #238
• Add support for CONCAT by @charlesbluca in #253
• [REVIEW] Fast path when possible for non numeric aggregation by @VibhuJawa in #236
• Restrict docker/deploy jobs to upstream repo, cancel concurrent test runs by @charlesbluca in #254
• Do not persist data to memory by default when creating tables by @jdye64 in #245
• Add flake8 pre-commit hook by @charlesbluca in #235
• Automatically label bugs / feature requests for triage by @charlesbluca in #261
• Support pandas style row udfs by @brandon-b-miller in #246
• Publish nightly builds to dask conda channel by @charlesbluca in #263
• Revert conda build tweaks by @charlesbluca in #266
• Try anaconda upload again for conda package upload by @charlesbluca in #267
• Feature/improve cli by @rajagurunath in #231
• Simplify DataContainer.assign operation by @charlesbluca in #271
• Added bug fix for window func by @rajagurunath in #277
• Pass return_type through to meta in apply by @brandon-b-miller in #275
• [Review] Add gpu tests for string functions by @ayushdg in #256
• Simplify single-partition sorting logic by @charlesbluca in #262
• Require UDF return type and update docs by @brandon-b-miller in #283

New Contributors

• @ayushdg made their first contribution in #220
• @charlesbluca made their first contribution in #244
• @quasiben made their first contribution in #229
• @VibhuJawa made their first contribution in #236
• @jdye64 made their first contribution in #245
• @brandon-b-miller made their first contribution in #246

Full Changelog: 0.3.9...0.4.0

Bugfixes

• Do not depend on pkg not specified in setup.py (#214)
• Use the mambaforge installer to speed up the build process (#216)
• Update all links from nils-braun to dask-contrib. (Fixes #212)
• Make JOINs also work for non-pandas dask dataframes (e.g. dask-cudf) (#211)

Bugfixes

• Remove mandatory dask_ml dependencies (#208)