Add a ScalarUDFImpl::simplfy() API, move SimplifyInfo et al to datafusion_expr

datafusion-examples/examples/expr_api.rs

@@ -24,15 +24,16 @@ use arrow::record_batch::RecordBatch;
 use datafusion::arrow::datatypes::{DataType, Field, Schema, TimeUnit};
 use datafusion::common::{DFField, DFSchema};
 use datafusion::error::Result;
-use datafusion::optimizer::simplify_expressions::{ExprSimplifier, SimplifyContext};
-use datafusion::physical_expr::execution_props::ExecutionProps;
+use datafusion::optimizer::simplify_expressions::ExprSimplifier;
 use datafusion::physical_expr::{
     analyze, create_physical_expr, AnalysisContext, ExprBoundaries, PhysicalExpr,
 };
 use datafusion::prelude::*;
 use datafusion_common::{ScalarValue, ToDFSchema};
+use datafusion_expr::execution_props::ExecutionProps;
 use datafusion_expr::expr::BinaryExpr;
 use datafusion_expr::interval_arithmetic::Interval;
+use datafusion_expr::simplify::SimplifyContext;
 use datafusion_expr::{ColumnarValue, ExprSchemable, Operator};
 
 /// This example demonstrates the DataFusion [`Expr`] API.

datafusion-examples/examples/simple_udtf.rs

@@ -28,8 +28,9 @@ use datafusion::physical_plan::memory::MemoryExec;
 use datafusion::physical_plan::ExecutionPlan;
 use datafusion::prelude::SessionContext;
 use datafusion_common::{plan_err, ScalarValue};
+use datafusion_expr::simplify::SimplifyContext;
 use datafusion_expr::{Expr, TableType};
-use datafusion_optimizer::simplify_expressions::{ExprSimplifier, SimplifyContext};
+use datafusion_optimizer::simplify_expressions::ExprSimplifier;
 use std::fs::File;
 use std::io::Seek;
 use std::path::Path;

datafusion/core/src/datasource/listing/helpers.rs

@@ -39,9 +39,9 @@ use crate::datasource::listing::ListingTableUrl;
 use crate::execution::context::SessionState;
 use datafusion_common::tree_node::{TreeNode, VisitRecursion};
 use datafusion_common::{internal_err, Column, DFField, DFSchema, DataFusionError};
+use datafusion_expr::execution_props::ExecutionProps;
 use datafusion_expr::{Expr, ScalarFunctionDefinition, Volatility};
 use datafusion_physical_expr::create_physical_expr;
-use datafusion_physical_expr::execution_props::ExecutionProps;
 use object_store::path::Path;
 use object_store::{ObjectMeta, ObjectStore};
 

datafusion/core/src/datasource/physical_plan/parquet/mod.rs

@@ -800,13 +800,14 @@ mod tests {
         ArrayRef, Date64Array, Int32Array, Int64Array, Int8Array, StringArray,
         StructArray,
     };
+
     use arrow::datatypes::{DataType, Field, Schema, SchemaBuilder};
     use arrow::record_batch::RecordBatch;
     use arrow_schema::Fields;
     use datafusion_common::{assert_contains, FileType, GetExt, ScalarValue, ToDFSchema};
+    use datafusion_expr::execution_props::ExecutionProps;
     use datafusion_expr::{col, lit, when, Expr};
     use datafusion_physical_expr::create_physical_expr;
-    use datafusion_physical_expr::execution_props::ExecutionProps;
 
     use chrono::{TimeZone, Utc};
     use futures::StreamExt;

datafusion/core/src/datasource/physical_plan/parquet/row_filter.rs

@@ -394,9 +394,9 @@ mod test {
     use super::*;
     use arrow::datatypes::Field;
     use datafusion_common::ToDFSchema;
+    use datafusion_expr::execution_props::ExecutionProps;
     use datafusion_expr::{cast, col, lit, Expr};
     use datafusion_physical_expr::create_physical_expr;
-    use datafusion_physical_expr::execution_props::ExecutionProps;
     use parquet::arrow::parquet_to_arrow_schema;
     use parquet::file::reader::{FileReader, SerializedFileReader};
     use rand::prelude::*;

datafusion/core/src/datasource/physical_plan/parquet/row_groups.rs

@@ -346,8 +346,8 @@ mod tests {
     use arrow::datatypes::Schema;
     use arrow::datatypes::{DataType, Field};
     use datafusion_common::{Result, ToDFSchema};
+    use datafusion_expr::execution_props::ExecutionProps;
     use datafusion_expr::{cast, col, lit, Expr};
-    use datafusion_physical_expr::execution_props::ExecutionProps;
     use datafusion_physical_expr::{create_physical_expr, PhysicalExpr};
     use parquet::arrow::arrow_to_parquet_schema;
     use parquet::arrow::async_reader::ParquetObjectReader;

datafusion/core/src/execution/context/mod.rs

@@ -43,12 +43,12 @@ use datafusion_common::{
     tree_node::{TreeNode, TreeNodeVisitor, VisitRecursion},
 };
 use datafusion_execution::registry::SerializerRegistry;
+pub use datafusion_expr::execution_props::ExecutionProps;
+use datafusion_expr::var_provider::is_system_variables;
 use datafusion_expr::{
     logical_plan::{DdlStatement, Statement},
     Expr, StringifiedPlan, UserDefinedLogicalNode, WindowUDF,
 };
-pub use datafusion_physical_expr::execution_props::ExecutionProps;
-use datafusion_physical_expr::var_provider::is_system_variables;
 use parking_lot::RwLock;
 use std::collections::hash_map::Entry;
 use std::string::String;

datafusion/core/src/physical_optimizer/pruning.rs

@@ -1338,10 +1338,10 @@ mod tests {
         datatypes::{DataType, TimeUnit},
     };
     use datafusion_common::{ScalarValue, ToDFSchema};
+    use datafusion_expr::execution_props::ExecutionProps;
     use datafusion_expr::expr::InList;
     use datafusion_expr::{cast, is_null, try_cast, Expr};
     use datafusion_physical_expr::create_physical_expr;
-    use datafusion_physical_expr::execution_props::ExecutionProps;
     use std::collections::HashMap;
     use std::ops::{Not, Rem};
 

datafusion/core/src/test_util/parquet.rs

@@ -28,9 +28,10 @@ use crate::datasource::listing::{ListingTableUrl, PartitionedFile};
 use crate::datasource::object_store::ObjectStoreUrl;
 use crate::datasource::physical_plan::{FileScanConfig, ParquetExec};
 use crate::error::Result;
-use crate::optimizer::simplify_expressions::{ExprSimplifier, SimplifyContext};
+use crate::logical_expr::execution_props::ExecutionProps;
+use crate::logical_expr::simplify::SimplifyContext;
+use crate::optimizer::simplify_expressions::ExprSimplifier;
 use crate::physical_expr::create_physical_expr;
-use crate::physical_expr::execution_props::ExecutionProps;
 use crate::physical_plan::filter::FilterExec;
 use crate::physical_plan::metrics::MetricsSet;
 use crate::physical_plan::ExecutionPlan;

datafusion/core/src/variable/mod.rs

@@ -17,4 +17,4 @@
 
 //! Variable provider for `@name` and `@@name` style runtime values.
 
-pub use datafusion_physical_expr::var_provider::{VarProvider, VarType};
+pub use datafusion_expr::var_provider::{VarProvider, VarType};

datafusion/core/tests/dataframe/mod.rs

@@ -48,13 +48,13 @@ use datafusion_common::{assert_contains, DataFusionError, ScalarValue, UnnestOpt
 use datafusion_execution::config::SessionConfig;
 use datafusion_execution::runtime_env::RuntimeEnv;
 use datafusion_expr::expr::{GroupingSet, Sort};
+use datafusion_expr::var_provider::{VarProvider, VarType};
 use datafusion_expr::{
     array_agg, avg, cast, col, count, exists, expr, in_subquery, lit, max, out_ref_col,
     placeholder, scalar_subquery, sum, when, wildcard, AggregateFunction, Expr,
     ExprSchemable, WindowFrame, WindowFrameBound, WindowFrameUnits,
     WindowFunctionDefinition,
 };
-use datafusion_physical_expr::var_provider::{VarProvider, VarType};
 
 #[tokio::test]
 async fn test_count_wildcard_on_sort() -> Result<()> {

datafusion/core/tests/parquet/page_pruning.rs

@@ -28,9 +28,9 @@ use datafusion::physical_plan::metrics::MetricValue;
 use datafusion::physical_plan::ExecutionPlan;
 use datafusion::prelude::SessionContext;
 use datafusion_common::{ScalarValue, Statistics, ToDFSchema};
+use datafusion_expr::execution_props::ExecutionProps;
 use datafusion_expr::{col, lit, Expr};
 use datafusion_physical_expr::create_physical_expr;
-use datafusion_physical_expr::execution_props::ExecutionProps;
 
 use futures::StreamExt;
 use object_store::path::Path;

datafusion/core/tests/simplification.rs

@@ -20,18 +20,17 @@
 use arrow::datatypes::{DataType, Field, Schema};
 use arrow_array::{ArrayRef, Int32Array};
 use chrono::{DateTime, TimeZone, Utc};
-use datafusion::common::DFSchema;
 use datafusion::{error::Result, execution::context::ExecutionProps, prelude::*};
 use datafusion_common::cast::as_int32_array;
 use datafusion_common::ScalarValue;
+use datafusion_common::{DFSchemaRef, ToDFSchema};
 use datafusion_expr::expr::ScalarFunction;
+use datafusion_expr::simplify::SimplifyInfo;
 use datafusion_expr::{
     expr, table_scan, BuiltinScalarFunction, Cast, ColumnarValue, Expr, ExprSchemable,
     LogicalPlan, LogicalPlanBuilder, ScalarUDF, Volatility,
 };
-use datafusion_optimizer::simplify_expressions::{
-    ExprSimplifier, SimplifyExpressions, SimplifyInfo,
-};
+use datafusion_optimizer::simplify_expressions::{ExprSimplifier, SimplifyExpressions};
 use datafusion_optimizer::{OptimizerContext, OptimizerRule};
 use std::sync::Arc;
 
@@ -42,7 +41,7 @@ use std::sync::Arc;
 /// objects or from some other implementation
 struct MyInfo {
     /// The input schema
-    schema: DFSchema,
+    schema: DFSchemaRef,
 
     /// Execution specific details needed for constant evaluation such
     /// as the current time for `now()` and [VariableProviders]
@@ -51,24 +50,27 @@ struct MyInfo {
 
 impl SimplifyInfo for MyInfo {
     fn is_boolean_type(&self, expr: &Expr) -> Result<bool> {
-        Ok(matches!(expr.get_type(&self.schema)?, DataType::Boolean))
+        Ok(matches!(
+            expr.get_type(self.schema.as_ref())?,
+            DataType::Boolean
+        ))
     }
 
     fn nullable(&self, expr: &Expr) -> Result<bool> {
-        expr.nullable(&self.schema)
+        expr.nullable(self.schema.as_ref())
     }
 
     fn execution_props(&self) -> &ExecutionProps {
         &self.execution_props
     }
 
     fn get_data_type(&self, expr: &Expr) -> Result<DataType> {
-        expr.get_type(&self.schema)
+        expr.get_type(self.schema.as_ref())
     }
 }
 
-impl From<DFSchema> for MyInfo {
-    fn from(schema: DFSchema) -> Self {
+impl From<DFSchemaRef> for MyInfo {
+    fn from(schema: DFSchemaRef) -> Self {
         Self {
             schema,
             execution_props: ExecutionProps::new(),
@@ -81,13 +83,13 @@ impl From<DFSchema> for MyInfo {
 /// a: Int32 (possibly with nulls)
 /// b: Int32
 /// s: Utf8
-fn schema() -> DFSchema {
+fn schema() -> DFSchemaRef {
     Schema::new(vec![
         Field::new("a", DataType::Int32, true),
         Field::new("b", DataType::Int32, false),
         Field::new("s", DataType::Utf8, false),
     ])
-    .try_into()
+    .to_dfschema_ref()
     .unwrap()
 }
 

datafusion/core/tests/user_defined/user_defined_scalar_functions.rs

@@ -16,7 +16,9 @@
 // under the License.
 
 use arrow::compute::kernels::numeric::add;
-use arrow_array::{Array, ArrayRef, Float64Array, Int32Array, RecordBatch, UInt8Array};
+use arrow_array::{
+    Array, ArrayRef, Float32Array, Float64Array, Int32Array, RecordBatch, UInt8Array,
+};
 use arrow_schema::DataType::Float64;
 use arrow_schema::{DataType, Field, Schema};
 use datafusion::prelude::*;
@@ -26,10 +28,13 @@ use datafusion_common::{
     assert_batches_eq, assert_batches_sorted_eq, cast::as_int32_array, not_impl_err,
     plan_err, ExprSchema, Result, ScalarValue,
 };
+use datafusion_expr::simplify::ExprSimplifyResult;
+use datafusion_expr::simplify::SimplifyInfo;
 use datafusion_expr::{
     create_udaf, create_udf, Accumulator, ColumnarValue, ExprSchemable,
     LogicalPlanBuilder, ScalarUDF, ScalarUDFImpl, Signature, Volatility,
 };
+
 use rand::{thread_rng, Rng};
 use std::any::Any;
 use std::iter;
@@ -514,6 +519,97 @@ async fn deregister_udf() -> Result<()> {
     Ok(())
 }
 
+#[derive(Debug)]
+struct CastToI64UDF {
+    signature: Signature,
+}
+
+impl CastToI64UDF {
+    fn new() -> Self {
+        Self {
+            signature: Signature::any(1, Volatility::Immutable),
+        }
+    }
+}
+
+impl ScalarUDFImpl for CastToI64UDF {
+    fn as_any(&self) -> &dyn Any {
+        self
+    }
+    fn name(&self) -> &str {
+        "cast_to_i64"
+    }
+    fn signature(&self) -> &Signature {
+        &self.signature
+    }
+    fn return_type(&self, _args: &[DataType]) -> Result<DataType> {
+        Ok(DataType::Int64)
+    }
+    // Wrap with Expr::Cast() to Int64
+    fn simplify(
+        &self,
+        mut args: Vec<Expr>,
+        info: &dyn SimplifyInfo,
+    ) -> Result<ExprSimplifyResult> {
+        // Note that Expr::cast_to requires an ExprSchema but simplify gets a
+        // SimplifyInfo so we have to replicate some of the casting logic here.
+        let source_type = info.get_data_type(&args[0])?;
+        if source_type == DataType::Int64 {
+            Ok(ExprSimplifyResult::Original(args))
+        } else {
+            // DataFusion should have ensured the function is called with just a
+            // single argument
+            assert_eq!(args.len(), 1);
+            let e = args.pop().unwrap();
+            Ok(ExprSimplifyResult::Simplified(Expr::Cast(
+                datafusion_expr::Cast {
+                    expr: Box::new(e),
+                    data_type: DataType::Int64,
+                },
+            )))
+        }
+    }
+    // Casting should be done in `simplify`, so we just return the first argument
+    fn invoke(&self, args: &[ColumnarValue]) -> Result<ColumnarValue> {
+        assert_eq!(args.len(), 1);
+        Ok(args.first().unwrap().clone())
+    }
+}
+
+#[tokio::test]
+async fn test_user_defined_functions_cast_to_i64() -> Result<()> {
+    let ctx = SessionContext::new();
+
+    let schema = Arc::new(Schema::new(vec![Field::new("x", DataType::Float32, false)]));
+
+    let batch = RecordBatch::try_new(
+        schema,
+        vec![Arc::new(Float32Array::from(vec![1.0, 2.0, 3.0]))],
+    )?;
+
+    ctx.register_batch("t", batch)?;
+
+    let cast_to_i64_udf = ScalarUDF::from(CastToI64UDF::new());
+    ctx.register_udf(cast_to_i64_udf);
+
+    let result = plan_and_collect(&ctx, "SELECT cast_to_i64(x) FROM t").await?;
+
+    assert_batches_eq!(
+        &[
+            "+------------------+",
+            "| cast_to_i64(t.x) |",
+            "+------------------+",
+            "| 1                |",
+            "| 2                |",
+            "| 3                |",
+            "+------------------+"
+        ],
+        &result
+    );
+
+    Ok(())
+}
+
 #[derive(Debug)]
 struct TakeUDF {
     signature: Signature,

datafusion/expr/Cargo.toml

@@ -40,6 +40,7 @@ ahash = { version = "0.8", default-features = false, features = [
 ] }
 arrow = { workspace = true }
 arrow-array = { workspace = true }
+chrono = { workspace = true }
 datafusion-common = { workspace = true, default-features = true }
 paste = "^1.0"
 sqlparser = { workspace = true }

datafusion/physical-expr/src/execution_props.rs → datafusion/expr/src/execution_props.rs

@@ -24,14 +24,12 @@ use std::sync::Arc;
 /// Holds per-query execution properties and data (such as statement
 /// starting timestamps).
 ///
-/// An [`ExecutionProps`] is created each time a [`LogicalPlan`] is
+/// An [`ExecutionProps`] is created each time a `LogicalPlan` is
 /// prepared for execution (optimized). If the same plan is optimized
 /// multiple times, a new `ExecutionProps` is created each time.
 ///
 /// It is important that this structure be cheap to create as it is
 /// done so during predicate pruning and expression simplification
-///
-/// [`LogicalPlan`]: datafusion_expr::LogicalPlan
 #[derive(Clone, Debug)]
 pub struct ExecutionProps {
     pub query_execution_start_time: DateTime<Utc>,