datafusion_physical_expr/
planner.rs

1// Licensed to the Apache Software Foundation (ASF) under one
2// or more contributor license agreements.  See the NOTICE file
3// distributed with this work for additional information
4// regarding copyright ownership.  The ASF licenses this file
5// to you under the Apache License, Version 2.0 (the
6// "License"); you may not use this file except in compliance
7// with the License.  You may obtain a copy of the License at
8//
9//   http://www.apache.org/licenses/LICENSE-2.0
10//
11// Unless required by applicable law or agreed to in writing,
12// software distributed under the License is distributed on an
13// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
14// KIND, either express or implied.  See the License for the
15// specific language governing permissions and limitations
16// under the License.
17
18use std::sync::Arc;
19
20use crate::ScalarFunctionExpr;
21use crate::{
22    expressions::{self, binary, like, similar_to, Column, Literal},
23    PhysicalExpr,
24};
25
26use arrow::datatypes::Schema;
27use datafusion_common::config::ConfigOptions;
28use datafusion_common::{
29    exec_err, not_impl_err, plan_err, DFSchema, Result, ScalarValue, ToDFSchema,
30};
31use datafusion_expr::execution_props::ExecutionProps;
32use datafusion_expr::expr::{
33    Alias, Cast, FieldMetadata, InList, Placeholder, ScalarFunction,
34};
35use datafusion_expr::var_provider::is_system_variables;
36use datafusion_expr::var_provider::VarType;
37use datafusion_expr::{
38    binary_expr, lit, Between, BinaryExpr, Expr, Like, Operator, TryCast,
39};
40
41/// [PhysicalExpr] evaluate DataFusion expressions such as `A + 1`, or `CAST(c1
42/// AS int)`.
43///
44/// [PhysicalExpr] are the physical counterpart to [Expr] used in logical
45/// planning, and can be evaluated directly on a [RecordBatch]. They are
46/// normally created from [Expr] by a [PhysicalPlanner] and can be created
47/// directly using [create_physical_expr].
48///
49/// A Physical expression knows its type, nullability and how to evaluate itself.
50///
51/// [PhysicalPlanner]: https://docs.rs/datafusion/latest/datafusion/physical_planner/trait.PhysicalPlanner.html
52/// [RecordBatch]: https://docs.rs/arrow/latest/arrow/record_batch/struct.RecordBatch.html
53///
54/// # Example: Create `PhysicalExpr` from `Expr`
55/// ```
56/// # use arrow::datatypes::{DataType, Field, Schema};
57/// # use datafusion_common::DFSchema;
58/// # use datafusion_expr::{Expr, col, lit};
59/// # use datafusion_physical_expr::create_physical_expr;
60/// # use datafusion_expr::execution_props::ExecutionProps;
61/// // For a logical expression `a = 1`, we can create a physical expression
62/// let expr = col("a").eq(lit(1));
63/// // To create a PhysicalExpr we need 1. a schema
64/// let schema = Schema::new(vec![Field::new("a", DataType::Int32, true)]);
65/// let df_schema = DFSchema::try_from(schema).unwrap();
66/// // 2. ExecutionProps
67/// let props = ExecutionProps::new();
68/// // We can now create a PhysicalExpr:
69/// let physical_expr = create_physical_expr(&expr, &df_schema, &props).unwrap();
70/// ```
71///
72/// # Example: Executing a PhysicalExpr to obtain [ColumnarValue]
73/// ```
74/// # use std::sync::Arc;
75/// # use arrow::array::{cast::AsArray, BooleanArray, Int32Array, RecordBatch};
76/// # use arrow::datatypes::{DataType, Field, Schema};
77/// # use datafusion_common::{assert_batches_eq, DFSchema};
78/// # use datafusion_expr::{Expr, col, lit, ColumnarValue};
79/// # use datafusion_physical_expr::create_physical_expr;
80/// # use datafusion_expr::execution_props::ExecutionProps;
81/// # let expr = col("a").eq(lit(1));
82/// # let schema = Schema::new(vec![Field::new("a", DataType::Int32, true)]);
83/// # let df_schema = DFSchema::try_from(schema.clone()).unwrap();
84/// # let props = ExecutionProps::new();
85/// // Given a PhysicalExpr, for `a = 1` we can evaluate it against a RecordBatch like this:
86/// let physical_expr = create_physical_expr(&expr, &df_schema, &props).unwrap();
87/// // Input of [1,2,3]
88/// let input_batch = RecordBatch::try_from_iter(vec![
89///   ("a", Arc::new(Int32Array::from(vec![1, 2, 3])) as _)
90/// ]).unwrap();
91/// // The result is a ColumnarValue (either an Array or a Scalar)
92/// let result = physical_expr.evaluate(&input_batch).unwrap();
93/// // In this case, a BooleanArray with the result of the comparison
94/// let ColumnarValue::Array(arr) = result else {
95///  panic!("Expected an array")
96/// };
97/// assert_eq!(arr.as_boolean(), &BooleanArray::from(vec![true, false, false]));
98/// ```
99///
100/// [ColumnarValue]: datafusion_expr::ColumnarValue
101///
102/// Create a physical expression from a logical expression ([Expr]).
103///
104/// # Arguments
105///
106/// * `e` - The logical expression
107/// * `input_dfschema` - The DataFusion schema for the input, used to resolve `Column` references
108///   to qualified or unqualified fields by name.
109pub fn create_physical_expr(
110    e: &Expr,
111    input_dfschema: &DFSchema,
112    execution_props: &ExecutionProps,
113) -> Result<Arc<dyn PhysicalExpr>> {
114    let input_schema: &Schema = &input_dfschema.into();
115
116    match e {
117        Expr::Alias(Alias { expr, metadata, .. }) => {
118            if let Expr::Literal(v, prior_metadata) = expr.as_ref() {
119                let new_metadata = FieldMetadata::merge_options(
120                    prior_metadata.as_ref(),
121                    metadata.as_ref(),
122                );
123                Ok(Arc::new(Literal::new_with_metadata(
124                    v.clone(),
125                    new_metadata,
126                )))
127            } else {
128                Ok(create_physical_expr(expr, input_dfschema, execution_props)?)
129            }
130        }
131        Expr::Column(c) => {
132            let idx = input_dfschema.index_of_column(c)?;
133            Ok(Arc::new(Column::new(&c.name, idx)))
134        }
135        Expr::Literal(value, metadata) => Ok(Arc::new(Literal::new_with_metadata(
136            value.clone(),
137            metadata.clone(),
138        ))),
139        Expr::ScalarVariable(_, variable_names) => {
140            if is_system_variables(variable_names) {
141                match execution_props.get_var_provider(VarType::System) {
142                    Some(provider) => {
143                        let scalar_value = provider.get_value(variable_names.clone())?;
144                        Ok(Arc::new(Literal::new(scalar_value)))
145                    }
146                    _ => plan_err!("No system variable provider found"),
147                }
148            } else {
149                match execution_props.get_var_provider(VarType::UserDefined) {
150                    Some(provider) => {
151                        let scalar_value = provider.get_value(variable_names.clone())?;
152                        Ok(Arc::new(Literal::new(scalar_value)))
153                    }
154                    _ => plan_err!("No user defined variable provider found"),
155                }
156            }
157        }
158        Expr::IsTrue(expr) => {
159            let binary_op = binary_expr(
160                expr.as_ref().clone(),
161                Operator::IsNotDistinctFrom,
162                lit(true),
163            );
164            create_physical_expr(&binary_op, input_dfschema, execution_props)
165        }
166        Expr::IsNotTrue(expr) => {
167            let binary_op =
168                binary_expr(expr.as_ref().clone(), Operator::IsDistinctFrom, lit(true));
169            create_physical_expr(&binary_op, input_dfschema, execution_props)
170        }
171        Expr::IsFalse(expr) => {
172            let binary_op = binary_expr(
173                expr.as_ref().clone(),
174                Operator::IsNotDistinctFrom,
175                lit(false),
176            );
177            create_physical_expr(&binary_op, input_dfschema, execution_props)
178        }
179        Expr::IsNotFalse(expr) => {
180            let binary_op =
181                binary_expr(expr.as_ref().clone(), Operator::IsDistinctFrom, lit(false));
182            create_physical_expr(&binary_op, input_dfschema, execution_props)
183        }
184        Expr::IsUnknown(expr) => {
185            let binary_op = binary_expr(
186                expr.as_ref().clone(),
187                Operator::IsNotDistinctFrom,
188                Expr::Literal(ScalarValue::Boolean(None), None),
189            );
190            create_physical_expr(&binary_op, input_dfschema, execution_props)
191        }
192        Expr::IsNotUnknown(expr) => {
193            let binary_op = binary_expr(
194                expr.as_ref().clone(),
195                Operator::IsDistinctFrom,
196                Expr::Literal(ScalarValue::Boolean(None), None),
197            );
198            create_physical_expr(&binary_op, input_dfschema, execution_props)
199        }
200        Expr::BinaryExpr(BinaryExpr { left, op, right }) => {
201            // Create physical expressions for left and right operands
202            let lhs = create_physical_expr(left, input_dfschema, execution_props)?;
203            let rhs = create_physical_expr(right, input_dfschema, execution_props)?;
204            // Note that the logical planner is responsible
205            // for type coercion on the arguments (e.g. if one
206            // argument was originally Int32 and one was
207            // Int64 they will both be coerced to Int64).
208            //
209            // There should be no coercion during physical
210            // planning.
211            binary(lhs, *op, rhs, input_schema)
212        }
213        Expr::Like(Like {
214            negated,
215            expr,
216            pattern,
217            escape_char,
218            case_insensitive,
219        }) => {
220            // `\` is the implicit escape, see https://github.com/apache/datafusion/issues/13291
221            if escape_char.unwrap_or('\\') != '\\' {
222                return exec_err!(
223                    "LIKE does not support escape_char other than the backslash (\\)"
224                );
225            }
226            let physical_expr =
227                create_physical_expr(expr, input_dfschema, execution_props)?;
228            let physical_pattern =
229                create_physical_expr(pattern, input_dfschema, execution_props)?;
230            like(
231                *negated,
232                *case_insensitive,
233                physical_expr,
234                physical_pattern,
235                input_schema,
236            )
237        }
238        Expr::SimilarTo(Like {
239            negated,
240            expr,
241            pattern,
242            escape_char,
243            case_insensitive,
244        }) => {
245            if escape_char.is_some() {
246                return exec_err!("SIMILAR TO does not support escape_char yet");
247            }
248            let physical_expr =
249                create_physical_expr(expr, input_dfschema, execution_props)?;
250            let physical_pattern =
251                create_physical_expr(pattern, input_dfschema, execution_props)?;
252            similar_to(*negated, *case_insensitive, physical_expr, physical_pattern)
253        }
254        Expr::Case(case) => {
255            let expr: Option<Arc<dyn PhysicalExpr>> = if let Some(e) = &case.expr {
256                Some(create_physical_expr(
257                    e.as_ref(),
258                    input_dfschema,
259                    execution_props,
260                )?)
261            } else {
262                None
263            };
264            let (when_expr, then_expr): (Vec<&Expr>, Vec<&Expr>) = case
265                .when_then_expr
266                .iter()
267                .map(|(w, t)| (w.as_ref(), t.as_ref()))
268                .unzip();
269            let when_expr =
270                create_physical_exprs(when_expr, input_dfschema, execution_props)?;
271            let then_expr =
272                create_physical_exprs(then_expr, input_dfschema, execution_props)?;
273            let when_then_expr: Vec<(Arc<dyn PhysicalExpr>, Arc<dyn PhysicalExpr>)> =
274                when_expr
275                    .iter()
276                    .zip(then_expr.iter())
277                    .map(|(w, t)| (Arc::clone(w), Arc::clone(t)))
278                    .collect();
279            let else_expr: Option<Arc<dyn PhysicalExpr>> =
280                if let Some(e) = &case.else_expr {
281                    Some(create_physical_expr(
282                        e.as_ref(),
283                        input_dfschema,
284                        execution_props,
285                    )?)
286                } else {
287                    None
288                };
289            Ok(expressions::case(expr, when_then_expr, else_expr)?)
290        }
291        Expr::Cast(Cast { expr, data_type }) => expressions::cast(
292            create_physical_expr(expr, input_dfschema, execution_props)?,
293            input_schema,
294            data_type.clone(),
295        ),
296        Expr::TryCast(TryCast { expr, data_type }) => expressions::try_cast(
297            create_physical_expr(expr, input_dfschema, execution_props)?,
298            input_schema,
299            data_type.clone(),
300        ),
301        Expr::Not(expr) => {
302            expressions::not(create_physical_expr(expr, input_dfschema, execution_props)?)
303        }
304        Expr::Negative(expr) => expressions::negative(
305            create_physical_expr(expr, input_dfschema, execution_props)?,
306            input_schema,
307        ),
308        Expr::IsNull(expr) => expressions::is_null(create_physical_expr(
309            expr,
310            input_dfschema,
311            execution_props,
312        )?),
313        Expr::IsNotNull(expr) => expressions::is_not_null(create_physical_expr(
314            expr,
315            input_dfschema,
316            execution_props,
317        )?),
318        Expr::ScalarFunction(ScalarFunction { func, args }) => {
319            let physical_args =
320                create_physical_exprs(args, input_dfschema, execution_props)?;
321            let config_options = match execution_props.config_options.as_ref() {
322                Some(config_options) => Arc::clone(config_options),
323                None => Arc::new(ConfigOptions::default()),
324            };
325
326            Ok(Arc::new(ScalarFunctionExpr::try_new(
327                Arc::clone(func),
328                physical_args,
329                input_schema,
330                config_options,
331            )?))
332        }
333        Expr::Between(Between {
334            expr,
335            negated,
336            low,
337            high,
338        }) => {
339            let value_expr = create_physical_expr(expr, input_dfschema, execution_props)?;
340            let low_expr = create_physical_expr(low, input_dfschema, execution_props)?;
341            let high_expr = create_physical_expr(high, input_dfschema, execution_props)?;
342
343            // rewrite the between into the two binary operators
344            let binary_expr = binary(
345                binary(
346                    Arc::clone(&value_expr),
347                    Operator::GtEq,
348                    low_expr,
349                    input_schema,
350                )?,
351                Operator::And,
352                binary(
353                    Arc::clone(&value_expr),
354                    Operator::LtEq,
355                    high_expr,
356                    input_schema,
357                )?,
358                input_schema,
359            );
360
361            if *negated {
362                expressions::not(binary_expr?)
363            } else {
364                binary_expr
365            }
366        }
367        Expr::InList(InList {
368            expr,
369            list,
370            negated,
371        }) => match expr.as_ref() {
372            Expr::Literal(ScalarValue::Utf8(None), _) => {
373                Ok(expressions::lit(ScalarValue::Boolean(None)))
374            }
375            _ => {
376                let value_expr =
377                    create_physical_expr(expr, input_dfschema, execution_props)?;
378
379                let list_exprs =
380                    create_physical_exprs(list, input_dfschema, execution_props)?;
381                expressions::in_list(value_expr, list_exprs, negated, input_schema)
382            }
383        },
384        Expr::Placeholder(Placeholder { id, .. }) => {
385            exec_err!("Placeholder '{id}' was not provided a value for execution.")
386        }
387        other => {
388            not_impl_err!("Physical plan does not support logical expression {other:?}")
389        }
390    }
391}
392
393/// Create vector of Physical Expression from a vector of logical expression
394pub fn create_physical_exprs<'a, I>(
395    exprs: I,
396    input_dfschema: &DFSchema,
397    execution_props: &ExecutionProps,
398) -> Result<Vec<Arc<dyn PhysicalExpr>>>
399where
400    I: IntoIterator<Item = &'a Expr>,
401{
402    exprs
403        .into_iter()
404        .map(|expr| create_physical_expr(expr, input_dfschema, execution_props))
405        .collect()
406}
407
408/// Convert a logical expression to a physical expression (without any simplification, etc)
409pub fn logical2physical(expr: &Expr, schema: &Schema) -> Arc<dyn PhysicalExpr> {
410    let df_schema = schema.clone().to_dfschema().unwrap();
411    let execution_props = ExecutionProps::new();
412    create_physical_expr(expr, &df_schema, &execution_props).unwrap()
413}
414
415#[cfg(test)]
416mod tests {
417    use arrow::array::{ArrayRef, BooleanArray, RecordBatch, StringArray};
418    use arrow::datatypes::{DataType, Field};
419
420    use datafusion_expr::{col, lit};
421
422    use super::*;
423
424    #[test]
425    fn test_create_physical_expr_scalar_input_output() -> Result<()> {
426        let expr = col("letter").eq(lit("A"));
427
428        let schema = Schema::new(vec![Field::new("letter", DataType::Utf8, false)]);
429        let df_schema = DFSchema::try_from_qualified_schema("data", &schema)?;
430        let p = create_physical_expr(&expr, &df_schema, &ExecutionProps::new())?;
431
432        let batch = RecordBatch::try_new(
433            Arc::new(schema),
434            vec![Arc::new(StringArray::from_iter_values(vec![
435                "A", "B", "C", "D",
436            ]))],
437        )?;
438        let result = p.evaluate(&batch)?;
439        let result = result.into_array(4).expect("Failed to convert to array");
440
441        assert_eq!(
442            &result,
443            &(Arc::new(BooleanArray::from(vec![true, false, false, false,])) as ArrayRef)
444        );
445
446        Ok(())
447    }
448}
datafusion_physical_expr/planner.rs

datafusion_physical_expr/
planner.rs
