feat: enable pickling for Python aggregate and window UDFs (#1545)
* feat: inline encoding for Python aggregate and window UDFs
Extends the PythonLogicalCodec / PythonPhysicalCodec inline encoding
introduced for scalar UDFs to also cover Python-defined aggregate and
window UDFs. The cloudpickle tuple shape per family is:
DFPYUDA (agg) (name, accumulator_factory, input_schema_bytes,
return_schema_bytes, state_schema_bytes,
volatility_str)
DFPYUDW (window) (name, evaluator_factory, input_schema_bytes,
return_schema_bytes, volatility_str)
Same wire-framing as scalar (family magic + version byte + cloudpickle
blob), same schema serde (arrow-rs native IPC), same cached cloudpickle
handle. The agg state schema is encoded as a full IPC schema so the
post-decode UDF reports the same names + nullability + metadata as the
sender — relevant for accumulators whose StateFieldsArgs consumers key
off names rather than positional DataType.
Required restructuring two existing UDF impls so the codec can grab
the Python callable directly:
* udaf.rs: replaces create_udaf + AccumulatorFactoryFunction closure
with a named PythonFunctionAggregateUDF that stores the Py<PyAny>
accumulator factory. Synthesizes state_{i} field names when the
Python constructor passes only Vec<DataType>; from_parts preserves
the full state schema on the decode side.
* udwf.rs: renames MultiColumnWindowUDF -> PythonFunctionWindowUDF,
drops the PartitionEvaluatorFactory PtrEq wrapper, stores the
Py<PyAny> evaluator directly. PartialEq and Hash get the same
pointer-identity fast path + debug-log exception handling already
on PythonFunctionScalarUDF.
User-facing surface:
* AggregateUDF.name and WindowUDF.name properties (parallel to the
ScalarUDF.name shipped in PR1).
* Existing UDAF/UDWF construction paths are unchanged.
The per-session with_python_udf_inlining toggle, sender-side context,
strict refusal, and user-guide docs land in PRs 3-4 of this series.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* feat: restore pub UDAF/UDWF helpers and document inline encoding
Re-export `to_rust_accumulator`, `to_rust_partition_evaluator`, and
`PythonFunctionWindowUDF` (with a `MultiColumnWindowUDF` alias) by
promoting `udaf` and `udwf` to `pub mod` so prior downstream Rust
consumers keep their API surface after the inline-encoding refactor.
Adds an end-to-end window UDF pickle round-trip test that runs the
decoded evaluator over a real session, mirroring the aggregate test.
Documents the cloudpickle-based shipping behavior of Python aggregate
and window UDFs in the user-guide aggregations and windows pages.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix: address PR #1545 review feedback
- Fix CountAcc.merge in pickle test: sum over states[0] (partition
counts), not over the list of state fields. The prior implementation
only added partition 0's count when merging across partitions.
- Drive test_agg_udf_evaluates_after_roundtrip with a two-batch
DataFrame so merge actually runs and the round-tripped state-field
schema is exercised end-to-end.
- Correct PY_AGG_UDF_FAMILY / PY_WINDOW_UDF_FAMILY doc comments and the
aggregate block comment to reference "return schema bytes" rather
than "return type" / "return_type_bytes" so the docs match the actual
on-wire layout.
- Keep `udaf` and `udwf` modules private (matching `udf`) and
selectively re-export the helpers downstream Rust consumers rely on
(`to_rust_accumulator`, `to_rust_partition_evaluator`,
`PythonFunctionWindowUDF`, `MultiColumnWindowUDF`) instead of
exposing the whole module surface.
- Rename codec helpers `*_agg_udf` -> `*_udaf` and `*_window_udf` ->
`*_udwf` for naming consistency with the Python public aliases.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This is a Python library that binds to Apache Arrow in-memory query engine DataFusion.
DataFusion's Python bindings can be used as a foundation for building new data systems in Python. Here are some examples:
For tips on tuning parallelism, see Maximizing CPU Usage in the configuration guide.
The following example demonstrates running a SQL query against a Parquet file using DataFusion, storing the results in a Pandas DataFrame, and then plotting a chart.
The Parquet file used in this example can be downloaded from the following page:
from datafusion import SessionContext # Create a DataFusion context ctx = SessionContext() # Register table with context ctx.register_parquet('taxi', 'yellow_tripdata_2021-01.parquet') # Execute SQL df = ctx.sql("select passenger_count, count(*) " "from taxi " "where passenger_count is not null " "group by passenger_count " "order by passenger_count") # convert to Pandas pandas_df = df.to_pandas() # create a chart fig = pandas_df.plot(kind="bar", title="Trip Count by Number of Passengers").get_figure() fig.savefig('chart.png')
This produces the following chart:
You can use SessionContext's register_view method to convert a DataFrame into a view and register it with the context.
from datafusion import SessionContext, col, literal # Create a DataFusion context ctx = SessionContext() # Create sample data data = {"a": [1, 2, 3, 4, 5], "b": [10, 20, 30, 40, 50]} # Create a DataFrame from the dictionary df = ctx.from_pydict(data, "my_table") # Filter the DataFrame (for example, keep rows where a > 2) df_filtered = df.filter(col("a") > literal(2)) # Register the dataframe as a view with the context ctx.register_view("view1", df_filtered) # Now run a SQL query against the registered view df_view = ctx.sql("SELECT * FROM view1") # Collect the results results = df_view.collect() # Convert results to a list of dictionaries for display result_dicts = [batch.to_pydict() for batch in results] print(result_dicts)
This will output:
[{'a': [3, 4, 5], 'b': [30, 40, 50]}]
It is possible to configure runtime (memory and disk settings) and configuration settings when creating a context.
runtime = ( RuntimeEnvBuilder() .with_disk_manager_os() .with_fair_spill_pool(10000000) ) config = ( SessionConfig() .with_create_default_catalog_and_schema(True) .with_default_catalog_and_schema("foo", "bar") .with_target_partitions(8) .with_information_schema(True) .with_repartition_joins(False) .with_repartition_aggregations(False) .with_repartition_windows(False) .with_parquet_pruning(False) .set("datafusion.execution.parquet.pushdown_filters", "true") ) ctx = SessionContext(config, runtime)
Refer to the API documentation for more information.
Printing the context will show the current configuration settings.
print(ctx)
For information about how to extend DataFusion Python, please see the extensions page of the online documentation.
See examples for more information.
uv add datafusion
pip install datafusion # or python -m pip install datafusion
conda install -c conda-forge datafusion
You can verify the installation by running:
>>> import datafusion >>> datafusion.__version__ '0.6.0'
This project ships a SKILL.md that teaches AI coding assistants how to write idiomatic DataFusion Python. It follows the Agent Skills open standard.
Preferred: npx skills add apache/datafusion-python — installs the skill in Claude Code, Cursor, Windsurf, Cline, Codex, Copilot, Gemini CLI, and other supported agents.
Manual: paste this line into your project's AGENTS.md / CLAUDE.md:
For DataFusion Python code, see https://github.com/apache/datafusion-python/blob/main/skills/datafusion_python/SKILL.md
This assumes that you have rust and cargo installed. We use the workflow recommended by pyo3 and maturin. The Maturin tools used in this workflow can be installed either via uv or pip. Both approaches should offer the same experience. It is recommended to use uv since it has significant performance improvements over pip.
Currently for protobuf support either protobuf or cmake must be installed.
Bootstrap (uv):
By default uv will attempt to build the datafusion python package. For our development we prefer to build manually. This means that when creating your virtual environment using uv sync you need to pass in the additional --no-install-package datafusion and for uv run commands the additional parameter --no-project
# fetch this repo git clone git@github.com:apache/datafusion-python.git # cd to the repo root cd datafusion-python/ # create the virtual environment uv sync --dev --no-install-package datafusion # activate the environment source .venv/bin/activate
Bootstrap (pip):
# fetch this repo git clone git@github.com:apache/datafusion-python.git # cd to the repo root cd datafusion-python/ # prepare development environment (used to build wheel / install in development) python3 -m venv .venv # activate the venv source .venv/bin/activate # update pip itself if necessary python -m pip install -U pip # install dependencies python -m pip install -r pyproject.toml
The tests rely on test data in git submodules.
git submodule update --init
Whenever rust code changes (your changes or via git pull):
# make sure you activate the venv using "source venv/bin/activate" first maturin develop --uv python -m pytest
Alternatively if you are using uv you can do the following without needing to activate the virtual environment:
uv run --no-project maturin develop --uv uv run --no-project pytest
To run the FFI tests within the examples folder, after you have built datafusion-python with the previous commands:
cd examples/datafusion-ffi-example uv run --no-project maturin develop --uv uv run --no-project pytest python/tests/_test_*py
datafusion-python takes advantage of pre-commit to assist developers with code linting to help reduce the number of commits that ultimately fail in CI due to linter errors. Using the pre-commit hooks is optional for the developer but certainly helpful for keeping PRs clean and concise.
Our pre-commit hooks can be installed by running pre-commit install, which will install the configurations in your DATAFUSION_PYTHON_ROOT/.github directory and run each time you perform a commit, failing to complete the commit if an offending lint is found allowing you to make changes locally before pushing.
The pre-commit hooks can also be run adhoc without installing them by simply running pre-commit run --all-files.
NOTE: the current pre-commit hooks require docker, and cmake. See note on protobuf above.
There are scripts in ci/scripts for running Rust and Python linters.
./ci/scripts/python_lint.sh ./ci/scripts/rust_clippy.sh ./ci/scripts/rust_fmt.sh ./ci/scripts/rust_toml_fmt.sh
This project includes an AI agent skill for auditing which features from the upstream Apache DataFusion Rust library are not yet exposed in these Python bindings. This is useful when adding missing functions, auditing API coverage, or ensuring parity with upstream.
The skill accepts an optional area argument:
scalar functions aggregate functions window functions dataframe session context ffi types all
If no argument is provided, it defaults to checking all areas. The skill will fetch the upstream DataFusion documentation, compare it against the functions and methods exposed in this project, and produce a coverage report listing what is currently exposed and what is missing.
The skill definition lives in .ai/skills/check-upstream/SKILL.md and follows the Agent Skills open standard. It can be used by any AI coding agent that supports skill discovery, or followed manually.
To change test dependencies, change the pyproject.toml and run
uv sync --dev --no-install-package datafusion