2023.4.0
What's Changed
- Minor doc fix to use use latest config option by @ayushdg in #1032
- Break off source distribution publication to separate release job by @charlesbluca in #1030
- Remove Dask constraint for development by @charlesbluca in #1034
- Rerun
test_dask_fsql
on Python 3.8 pytests by @sarahyurick in #1021 - Bump docker/build-push-action from 3 to 4 by @dependabot in #1028
- Bump pyo3 from 0.18.0 to 0.18.1 in /dask_planner by @dependabot in #1035
- CI improvements by @ayushdg in #1033
- Add best practices doc by @ChrisJar in #967
- Upgrade to DataFusion 17.0.0 by @andygrove in #998
- Table Statistics Support by @sarahyurick in #1037
- Run GPU testing against python 3.8/3.10 by @charlesbluca in #1042
- Update gpuCI
RAPIDS_VER
to23.04
by @github-actions in #1024 - Remove call to
nest_asyncio.apply
in server startup by @charlesbluca in #1048 - Remove
nest-asyncio
from dependencies by @charlesbluca in #1051 - Add is_distributed implementation for fugue DaskSQLEngine by @ayushdg in #1059
- [REVIEW] Bump to datafusion-18 by @jdye64 in #1046
- Split gpuCI environment into 3.8/3.10 environments by @charlesbluca in #1052
- Resolve tests in
test_server.py
by @charlesbluca in #1049 - Fix CI failures by @ayushdg in #1067
- Increase number of
test_dask_fsql
reruns by @sarahyurick in #1075 - Support column names starting with a number by @sarahyurick in #1060
- Unblock P2P shuffle operations on CPU by @charlesbluca in #1076
- Add
filepath
fields in Python and Rust by @sarahyurick in #1074 - Disable dependabot for
datafusion
major updates,cibuildwheel
patch updates by @charlesbluca in #1077 - Fix getitem bug in parquet_statistics by @rjzamora in #1072
- Pin
sqlalchemy<2
in CI environments by @charlesbluca in #1080 - Add pytest for DataContainer statistics by @sarahyurick in #1061
- Add server support for null values by @charlesbluca in #1062
- Bump actions/upload-artifact from 1 to 3 by @dependabot in #1065
- Bump jacobtomlinson/gha-find-replace from 2 to 3 by @dependabot in #1039
- Update machine learning documentation by @sarahyurick in #1043
- Remove @andygrove from Rust codeowners by @charlesbluca in #1086
- Bump async-trait from 0.1.64 to 0.1.66 in /dask_planner by @dependabot in #1071
- Fix clippy style errors by @ayushdg in #1087
- Install upstream dask with dependencies in import testing by @charlesbluca in #1091
- Bump tokio from 1.25.0 to 1.26.0 in /dask_planner by @dependabot in #1070
- Add columns data type creation arguments to server response for
CHAR
/VARCHAR
columns by @charlesbluca in #1089 - Bump async-trait from 0.1.66 to 0.1.68 in /dask_planner by @dependabot in #1096
- Bump pyo3 from 0.18.1 to 0.18.2 in /dask_planner by @dependabot in #1097
- Bump tokio from 1.26.0 to 1.27.0 in /dask_planner by @dependabot in #1098
- Add Rust -> Python logging bridge by @jdye64 in #1100
- Ignore case sensitivity of columns for case_sensitive=False by @ayushdg in #1090
- Remove deprecated
file_format
argument,register_dask_table
function by @charlesbluca in #1103 - Update gpuCI
RAPIDS_VER
to23.06
by @github-actions in #1105 - Remove
dask/gpu
team from gpuCI bump PR reviewers by @charlesbluca in #1107 - Bump mimalloc from 0.1.34 to 0.1.36 in /dask_planner by @charlesbluca in #1109
- Implement join reordering of fact-dimension joins by @sarahyurick in #1027
- Add support for optional catalog name to
SHOW SCHEMAS | TABLES
by @charlesbluca in #1063 - Pin dask/distributed for release by @charlesbluca in #1112
Full Changelog: 2023.2.0...2023.4.0