DataFusion 13.0.0

Cargo.toml

@@ -34,9 +34,9 @@ default = ["mimalloc"]
 tokio = { version = "1.0", features = ["macros", "rt", "rt-multi-thread", "sync"] }
 rand = "0.7"
 pyo3 = { version = "~0.17.1", features = ["extension-module", "abi3", "abi3-py37"] }
-datafusion = { version = "^12.0.0", features = ["pyarrow", "avro"] }
-datafusion-expr = { version = "^12.0.0" }
-datafusion-common = { version = "^12.0.0", features = ["pyarrow"] }
+datafusion = { version = "^13.0.0", features = ["pyarrow", "avro"] }
+datafusion-expr = { version = "^13.0.0" }
+datafusion-common = { version = "^13.0.0", features = ["pyarrow"] }
 uuid = { version = "0.8", features = ["v4"] }
 mimalloc = { version = "*", optional = true, default-features = false }
 async-trait = "0.1"
@@ -51,4 +51,4 @@ name = "datafusion._internal"
 
 [profile.release]
 lto = true
-codegen-units = 1
+codegen-units = 1

README.md

@@ -163,6 +163,13 @@ python -m pip install -U pip
 python -m pip install -r requirements-310.txt
 ```
 
+The tests rely on test data in git submodules.
+
+```bash
+git submodule init
+git submodule update
+```
+
 Whenever rust code changes (your changes or via `git pull`):
 
 ```bash

src/context.rs

@@ -24,6 +24,7 @@ use pyo3::exceptions::{PyKeyError, PyValueError};
 use pyo3::prelude::*;
 
 use datafusion::arrow::datatypes::Schema;
+use datafusion::arrow::pyarrow::PyArrowType;
 use datafusion::arrow::record_batch::RecordBatch;
 use datafusion::datasource::datasource::TableProvider;
 use datafusion::datasource::MemTable;
@@ -99,9 +100,12 @@ impl PySessionContext {
         Ok(PyDataFrame::new(df))
     }
 
-    fn create_dataframe(&mut self, partitions: Vec<Vec<RecordBatch>>) -> PyResult<PyDataFrame> {
-        let table = MemTable::try_new(partitions[0][0].schema(), partitions)
-            .map_err(DataFusionError::from)?;
+    fn create_dataframe(
+        &mut self,
+        partitions: PyArrowType<Vec<Vec<RecordBatch>>>,
+    ) -> PyResult<PyDataFrame> {
+        let schema = partitions.0[0][0].schema();
+        let table = MemTable::try_new(schema, partitions.0).map_err(DataFusionError::from)?;
 
         // generate a random (unique) name for this table
         // table name cannot start with numeric digit
@@ -136,10 +140,10 @@ impl PySessionContext {
     fn register_record_batches(
         &mut self,
         name: &str,
-        partitions: Vec<Vec<RecordBatch>>,
+        partitions: PyArrowType<Vec<Vec<RecordBatch>>>,
     ) -> PyResult<()> {
-        let schema = partitions[0][0].schema();
-        let table = MemTable::try_new(schema, partitions)?;
+        let schema = partitions.0[0][0].schema();
+        let table = MemTable::try_new(schema, partitions.0)?;
         self.ctx
             .register_table(name, Arc::new(table))
             .map_err(DataFusionError::from)?;
@@ -182,7 +186,7 @@ impl PySessionContext {
         &mut self,
         name: &str,
         path: PathBuf,
-        schema: Option<Schema>,
+        schema: Option<PyArrowType<Schema>>,
         has_header: bool,
         delimiter: &str,
         schema_infer_max_records: usize,
@@ -204,7 +208,7 @@ impl PySessionContext {
             .delimiter(delimiter[0])
             .schema_infer_max_records(schema_infer_max_records)
             .file_extension(file_extension);
-        options.schema = schema.as_ref();
+        options.schema = schema.as_ref().map(|x| &x.0);
 
         let result = self.ctx.register_csv(name, path, options);
         wait_for_future(py, result).map_err(DataFusionError::from)?;
@@ -277,7 +281,7 @@ impl PySessionContext {
     fn read_csv(
         &self,
         path: PathBuf,
-        schema: Option<Schema>,
+        schema: Option<PyArrowType<Schema>>,
         has_header: bool,
         delimiter: &str,
         schema_infer_max_records: usize,
@@ -302,12 +306,17 @@ impl PySessionContext {
             .schema_infer_max_records(schema_infer_max_records)
             .file_extension(file_extension)
             .table_partition_cols(table_partition_cols);
-        options.schema = schema.as_ref();
 
-        let result = self.ctx.read_csv(path, options);
-        let df = PyDataFrame::new(wait_for_future(py, result).map_err(DataFusionError::from)?);
-
-        Ok(df)
+        if let Some(py_schema) = schema {
+            options.schema = Some(&py_schema.0);
+            let result = self.ctx.read_csv(path, options);
+            let df = PyDataFrame::new(wait_for_future(py, result).map_err(DataFusionError::from)?);
+            Ok(df)
+        } else {
+            let result = self.ctx.read_csv(path, options);
+            let df = PyDataFrame::new(wait_for_future(py, result).map_err(DataFusionError::from)?);
+            Ok(df)
+        }
     }
 
     #[allow(clippy::too_many_arguments)]
@@ -346,14 +355,14 @@ impl PySessionContext {
     fn read_avro(
         &self,
         path: &str,
-        schema: Option<Schema>,
+        schema: Option<PyArrowType<Schema>>,
         table_partition_cols: Vec<String>,
         file_extension: &str,
         py: Python,
     ) -> PyResult<PyDataFrame> {
         let mut options = AvroReadOptions::default().table_partition_cols(table_partition_cols);
         options.file_extension = file_extension;
-        options.schema = schema.map(Arc::new);
+        options.schema = schema.map(|s| Arc::new(s.0));
 
         let result = self.ctx.read_avro(path, options);
         let df = PyDataFrame::new(wait_for_future(py, result).map_err(DataFusionError::from)?);

src/dataframe.rs

@@ -18,7 +18,7 @@
 use crate::utils::wait_for_future;
 use crate::{errors::DataFusionError, expression::PyExpr};
 use datafusion::arrow::datatypes::Schema;
-use datafusion::arrow::pyarrow::PyArrowConvert;
+use datafusion::arrow::pyarrow::{PyArrowConvert, PyArrowException, PyArrowType};
 use datafusion::arrow::util::pretty;
 use datafusion::dataframe::DataFrame;
 use datafusion::prelude::*;
@@ -65,8 +65,8 @@ impl PyDataFrame {
     }
 
     /// Returns the schema from the logical plan
-    fn schema(&self) -> Schema {
-        self.df.schema().into()
+    fn schema(&self) -> PyArrowType<Schema> {
+        PyArrowType(self.df.schema().into())
     }
 
     #[args(args = "*")]
@@ -144,7 +144,7 @@ impl PyDataFrame {
     fn show(&self, py: Python, num: usize) -> PyResult<()> {
         let df = self.df.limit(0, Some(num))?;
         let batches = wait_for_future(py, df.collect())?;
-        Ok(pretty::print_batches(&batches)?)
+        pretty::print_batches(&batches).map_err(|err| PyArrowException::new_err(err.to_string()))
     }
 
     /// Filter out duplicate rows
@@ -186,7 +186,7 @@ impl PyDataFrame {
     fn explain(&self, py: Python, verbose: bool, analyze: bool) -> PyResult<()> {
         let df = self.df.explain(verbose, analyze)?;
         let batches = wait_for_future(py, df.collect())?;
-        Ok(pretty::print_batches(&batches)?)
+        pretty::print_batches(&batches).map_err(|err| PyArrowException::new_err(err.to_string()))
     }
 
     /// Repartition a `DataFrame` based on a logical partitioning scheme.

src/dataset.rs

@@ -27,6 +27,7 @@ use std::sync::Arc;
 use async_trait::async_trait;
 
 use datafusion::arrow::datatypes::SchemaRef;
+use datafusion::arrow::pyarrow::PyArrowType;
 use datafusion::datasource::datasource::TableProviderFilterPushDown;
 use datafusion::datasource::{TableProvider, TableType};
 use datafusion::error::{DataFusionError, Result as DFResult};
@@ -74,7 +75,14 @@ impl TableProvider for Dataset {
         Python::with_gil(|py| {
             let dataset = self.dataset.as_ref(py);
             // This can panic but since we checked that self.dataset is a pyarrow.dataset.Dataset it should never
-            Arc::new(dataset.getattr("schema").unwrap().extract().unwrap())
+            Arc::new(
+                dataset
+                    .getattr("schema")
+                    .unwrap()
+                    .extract::<PyArrowType<_>>()
+                    .unwrap()
+                    .0,
+            )
         })
     }
 

src/dataset_exec.rs

@@ -28,6 +28,7 @@ use futures::stream;
 use datafusion::arrow::datatypes::SchemaRef;
 use datafusion::arrow::error::ArrowError;
 use datafusion::arrow::error::Result as ArrowResult;
+use datafusion::arrow::pyarrow::PyArrowType;
 use datafusion::arrow::record_batch::RecordBatch;
 use datafusion::error::{DataFusionError as InnerDataFusionError, Result as DFResult};
 use datafusion::execution::context::TaskContext;
@@ -54,7 +55,7 @@ impl Iterator for PyArrowBatchesAdapter {
             Some(
                 batches
                     .next()?
-                    .and_then(|batch| batch.extract())
+                    .and_then(|batch| Ok(batch.extract::<PyArrowType<_>>()?.0))
                     .map_err(|err| ArrowError::ExternalError(Box::new(err))),
             )
         })
@@ -109,7 +110,12 @@ impl DatasetExec {
 
         let scanner = dataset.call_method("scanner", (), Some(kwargs))?;
 
-        let schema = Arc::new(scanner.getattr("projected_schema")?.extract()?);
+        let schema = Arc::new(
+            scanner
+                .getattr("projected_schema")?
+                .extract::<PyArrowType<_>>()?
+                .0,
+        );
 
         let builtins = Python::import(py, "builtins")?;
         let pylist = builtins.getattr("list")?;
@@ -211,7 +217,7 @@ impl ExecutionPlan for DatasetExec {
             let schema: SchemaRef = Arc::new(
                 scanner
                     .getattr("projected_schema")
-                    .and_then(|schema| schema.extract())
+                    .and_then(|schema| Ok(schema.extract::<PyArrowType<_>>()?.0))
                     .map_err(|err| InnerDataFusionError::External(Box::new(err)))?,
             );
             let record_batches: &PyIterator = scanner

src/expression.rs

@@ -19,6 +19,7 @@ use pyo3::{basic::CompareOp, prelude::*};
 use std::convert::{From, Into};
 
 use datafusion::arrow::datatypes::DataType;
+use datafusion::arrow::pyarrow::PyArrowType;
 use datafusion_expr::{col, lit, Expr};
 
 use datafusion::scalar::ScalarValue;
@@ -125,12 +126,12 @@ impl PyExpr {
         self.expr.clone().is_null().into()
     }
 
-    pub fn cast(&self, to: DataType) -> PyExpr {
+    pub fn cast(&self, to: PyArrowType<DataType>) -> PyExpr {
         // self.expr.cast_to() requires DFSchema to validate that the cast
         // is supported, omit that for now
         let expr = Expr::Cast {
             expr: Box::new(self.expr.clone()),
-            data_type: to,
+            data_type: to.0,
         };
         expr.into()
     }

src/udaf.rs

@@ -19,9 +19,9 @@ use std::sync::Arc;
 
 use pyo3::{prelude::*, types::PyTuple};
 
-use datafusion::arrow::array::ArrayRef;
+use datafusion::arrow::array::{Array, ArrayRef};
 use datafusion::arrow::datatypes::DataType;
-use datafusion::arrow::pyarrow::PyArrowConvert;
+use datafusion::arrow::pyarrow::{PyArrowConvert, PyArrowType};
 use datafusion::common::ScalarValue;
 use datafusion::error::{DataFusionError, Result};
 use datafusion_expr::{
@@ -82,6 +82,7 @@ impl Accumulator for RustAccumulator {
 
             // 1. cast states to Pyarrow array
             let state = state
+                .data()
                 .to_pyarrow(py)
                 .map_err(|e| DataFusionError::Execution(format!("{}", e)))?;
 
@@ -120,18 +121,18 @@ impl PyAggregateUDF {
     fn new(
         name: &str,
         accumulator: PyObject,
-        input_type: DataType,
-        return_type: DataType,
-        state_type: Vec<DataType>,
+        input_type: PyArrowType<DataType>,
+        return_type: PyArrowType<DataType>,
+        state_type: PyArrowType<Vec<DataType>>,
         volatility: &str,
     ) -> PyResult<Self> {
         let function = create_udaf(
             name,
-            input_type,
-            Arc::new(return_type),
+            input_type.0,
+            Arc::new(return_type.0),
             parse_volatility(volatility)?,
             to_rust_accumulator(accumulator),
-            Arc::new(state_type),
+            Arc::new(state_type.0),
         );
         Ok(Self { function })
     }

src/udf.rs

@@ -19,9 +19,9 @@ use std::sync::Arc;
 
 use pyo3::{prelude::*, types::PyTuple};
 
-use datafusion::arrow::array::ArrayRef;
+use datafusion::arrow::array::{make_array, Array, ArrayData, ArrayRef};
 use datafusion::arrow::datatypes::DataType;
-use datafusion::arrow::pyarrow::PyArrowConvert;
+use datafusion::arrow::pyarrow::{PyArrowConvert, PyArrowType};
 use datafusion::error::DataFusionError;
 use datafusion::physical_plan::functions::make_scalar_function;
 use datafusion::physical_plan::udf::ScalarUDF;
@@ -46,15 +46,14 @@ fn to_rust_function(func: PyObject) -> ScalarFunctionImplementation {
                 let py_args = PyTuple::new(py, py_args);
 
                 // 2. call function
-                let value = func.as_ref(py).call(py_args, None);
-                let value = match value {
-                    Ok(n) => Ok(n),
-                    Err(error) => Err(DataFusionError::Execution(format!("{:?}", error))),
-                }?;
+                let value = func
+                    .as_ref(py)
+                    .call(py_args, None)
+                    .map_err(|e| DataFusionError::Execution(format!("{:?}", e)))?;
 
                 // 3. cast to arrow::array::Array
-                let array = ArrayRef::from_pyarrow(value).unwrap();
-                Ok(array)
+                let array_data = ArrayData::from_pyarrow(value).unwrap();
+                Ok(make_array(array_data))
             })
         },
     )
@@ -73,14 +72,14 @@ impl PyScalarUDF {
     fn new(
         name: &str,
         func: PyObject,
-        input_types: Vec<DataType>,
-        return_type: DataType,
+        input_types: PyArrowType<Vec<DataType>>,
+        return_type: PyArrowType<DataType>,
         volatility: &str,
     ) -> PyResult<Self> {
         let function = create_udf(
             name,
-            input_types,
-            Arc::new(return_type),
+            input_types.0,
+            Arc::new(return_type.0),
             parse_volatility(volatility)?,
             to_rust_function(func),
         );