datafusion_physical_expr_common/

physical_expr.rs

// Licensed to the Apache Software Foundation (ASF) under one
// or more contributor license agreements.  See the NOTICE file
// distributed with this work for additional information
// regarding copyright ownership.  The ASF licenses this file
// to you under the Apache License, Version 2.0 (the
// "License"); you may not use this file except in compliance
// with the License.  You may obtain a copy of the License at
//
//   http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing,
// software distributed under the License is distributed on an
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, either express or implied.  See the License for the
// specific language governing permissions and limitations
// under the License.

use std::any::Any;
use std::fmt;
use std::fmt::{Debug, Display, Formatter};
use std::hash::{Hash, Hasher};
use std::sync::Arc;

use crate::utils::scatter;

use arrow::array::BooleanArray;
use arrow::compute::filter_record_batch;
use arrow::datatypes::{DataType, Schema};
use arrow::record_batch::RecordBatch;
use datafusion_common::tree_node::{Transformed, TransformedResult, TreeNode};
use datafusion_common::{internal_err, not_impl_err, Result, ScalarValue};
use datafusion_expr_common::columnar_value::ColumnarValue;
use datafusion_expr_common::interval_arithmetic::Interval;
use datafusion_expr_common::sort_properties::ExprProperties;
use datafusion_expr_common::statistics::Distribution;

use itertools::izip;

/// Shared [`PhysicalExpr`].
pub type PhysicalExprRef = Arc<dyn PhysicalExpr>;

/// [`PhysicalExpr`]s represent expressions such as `A + 1` or `CAST(c1 AS int)`.
///
/// `PhysicalExpr` knows its type, nullability and can be evaluated directly on
/// a [`RecordBatch`] (see [`Self::evaluate`]).
///
/// `PhysicalExpr` are the physical counterpart to [`Expr`] used in logical
/// planning. They are typically created from [`Expr`] by a [`PhysicalPlanner`]
/// invoked from a higher level API
///
/// Some important examples of `PhysicalExpr` are:
/// * [`Column`]: Represents a column at a given index in a RecordBatch
///
/// To create `PhysicalExpr` from  `Expr`, see
/// * [`SessionContext::create_physical_expr`]: A high level API
/// * [`create_physical_expr`]: A low level API
///
/// # Formatting `PhysicalExpr` as strings
/// There are three ways to format `PhysicalExpr` as a string:
/// * [`Debug`]: Standard Rust debugging format (e.g. `Constant { value: ... }`)
/// * [`Display`]: Detailed SQL-like format that shows expression structure (e.g. (`Utf8 ("foobar")`). This is often used for debugging and tests
/// * [`Self::fmt_sql`]: SQL-like human readable format (e.g. ('foobar')`), See also [`sql_fmt`]
///
/// [`SessionContext::create_physical_expr`]: https://docs.rs/datafusion/latest/datafusion/execution/context/struct.SessionContext.html#method.create_physical_expr
/// [`PhysicalPlanner`]: https://docs.rs/datafusion/latest/datafusion/physical_planner/trait.PhysicalPlanner.html
/// [`Expr`]: https://docs.rs/datafusion/latest/datafusion/logical_expr/enum.Expr.html
/// [`create_physical_expr`]: https://docs.rs/datafusion/latest/datafusion/physical_expr/fn.create_physical_expr.html
/// [`Column`]: https://docs.rs/datafusion/latest/datafusion/physical_expr/expressions/struct.Column.html
pub trait PhysicalExpr: Send + Sync + Display + Debug + DynEq + DynHash {
    /// Returns the physical expression as [`Any`] so that it can be
    /// downcast to a specific implementation.
    fn as_any(&self) -> &dyn Any;
    /// Get the data type of this expression, given the schema of the input
    fn data_type(&self, input_schema: &Schema) -> Result<DataType>;
    /// Determine whether this expression is nullable, given the schema of the input
    fn nullable(&self, input_schema: &Schema) -> Result<bool>;
    /// Evaluate an expression against a RecordBatch
    fn evaluate(&self, batch: &RecordBatch) -> Result<ColumnarValue>;
    /// Evaluate an expression against a RecordBatch after first applying a
    /// validity array
    fn evaluate_selection(
        &self,
        batch: &RecordBatch,
        selection: &BooleanArray,
    ) -> Result<ColumnarValue> {
        let tmp_batch = filter_record_batch(batch, selection)?;

        let tmp_result = self.evaluate(&tmp_batch)?;

        if batch.num_rows() == tmp_batch.num_rows() {
            // All values from the `selection` filter are true.
            Ok(tmp_result)
        } else if let ColumnarValue::Array(a) = tmp_result {
            scatter(selection, a.as_ref()).map(ColumnarValue::Array)
        } else {
            Ok(tmp_result)
        }
    }

    /// Get a list of child PhysicalExpr that provide the input for this expr.
    fn children(&self) -> Vec<&Arc<dyn PhysicalExpr>>;

    /// Returns a new PhysicalExpr where all children were replaced by new exprs.
    fn with_new_children(
        self: Arc<Self>,
        children: Vec<Arc<dyn PhysicalExpr>>,
    ) -> Result<Arc<dyn PhysicalExpr>>;

    /// Computes the output interval for the expression, given the input
    /// intervals.
    ///
    /// # Parameters
    ///
    /// * `children` are the intervals for the children (inputs) of this
    ///   expression.
    ///
    /// # Returns
    ///
    /// A `Result` containing the output interval for the expression in
    /// case of success, or an error object in case of failure.
    ///
    /// # Example
    ///
    /// If the expression is `a + b`, and the input intervals are `a: [1, 2]`
    /// and `b: [3, 4]`, then the output interval would be `[4, 6]`.
    fn evaluate_bounds(&self, _children: &[&Interval]) -> Result<Interval> {
        not_impl_err!("Not implemented for {self}")
    }

    /// Updates bounds for child expressions, given a known interval for this
    /// expression.
    ///
    /// This is used to propagate constraints down through an expression tree.
    ///
    /// # Parameters
    ///
    /// * `interval` is the currently known interval for this expression.
    /// * `children` are the current intervals for the children of this expression.
    ///
    /// # Returns
    ///
    /// A `Result` containing a `Vec` of new intervals for the children (in order)
    /// in case of success, or an error object in case of failure.
    ///
    /// If constraint propagation reveals an infeasibility for any child, returns
    /// [`None`]. If none of the children intervals change as a result of
    /// propagation, may return an empty vector instead of cloning `children`.
    /// This is the default (and conservative) return value.
    ///
    /// # Example
    ///
    /// If the expression is `a + b`, the current `interval` is `[4, 5]` and the
    /// inputs `a` and `b` are respectively given as `[0, 2]` and `[-∞, 4]`, then
    /// propagation would return `[0, 2]` and `[2, 4]` as `b` must be at least
    /// `2` to make the output at least `4`.
    fn propagate_constraints(
        &self,
        _interval: &Interval,
        _children: &[&Interval],
    ) -> Result<Option<Vec<Interval>>> {
        Ok(Some(vec![]))
    }

    /// Computes the output statistics for the expression, given the input
    /// statistics.
    ///
    /// # Parameters
    ///
    /// * `children` are the statistics for the children (inputs) of this
    ///   expression.
    ///
    /// # Returns
    ///
    /// A `Result` containing the output statistics for the expression in
    /// case of success, or an error object in case of failure.
    ///
    /// Expressions (should) implement this function and utilize the independence
    /// assumption, match on children distribution types and compute the output
    /// statistics accordingly. The default implementation simply creates an
    /// unknown output distribution by combining input ranges. This logic loses
    /// distribution information, but is a safe default.
    fn evaluate_statistics(&self, children: &[&Distribution]) -> Result<Distribution> {
        let children_ranges = children
            .iter()
            .map(|c| c.range())
            .collect::<Result<Vec<_>>>()?;
        let children_ranges_refs = children_ranges.iter().collect::<Vec<_>>();
        let output_interval = self.evaluate_bounds(children_ranges_refs.as_slice())?;
        let dt = output_interval.data_type();
        if dt.eq(&DataType::Boolean) {
            let p = if output_interval.eq(&Interval::CERTAINLY_TRUE) {
                ScalarValue::new_one(&dt)
            } else if output_interval.eq(&Interval::CERTAINLY_FALSE) {
                ScalarValue::new_zero(&dt)
            } else {
                ScalarValue::try_from(&dt)
            }?;
            Distribution::new_bernoulli(p)
        } else {
            Distribution::new_from_interval(output_interval)
        }
    }

    /// Updates children statistics using the given parent statistic for this
    /// expression.
    ///
    /// This is used to propagate statistics down through an expression tree.
    ///
    /// # Parameters
    ///
    /// * `parent` is the currently known statistics for this expression.
    /// * `children` are the current statistics for the children of this expression.
    ///
    /// # Returns
    ///
    /// A `Result` containing a `Vec` of new statistics for the children (in order)
    /// in case of success, or an error object in case of failure.
    ///
    /// If statistics propagation reveals an infeasibility for any child, returns
    /// [`None`]. If none of the children statistics change as a result of
    /// propagation, may return an empty vector instead of cloning `children`.
    /// This is the default (and conservative) return value.
    ///
    /// Expressions (should) implement this function and apply Bayes rule to
    /// reconcile and update parent/children statistics. This involves utilizing
    /// the independence assumption, and matching on distribution types. The
    /// default implementation simply creates an unknown distribution if it can
    /// narrow the range by propagating ranges. This logic loses distribution
    /// information, but is a safe default.
    fn propagate_statistics(
        &self,
        parent: &Distribution,
        children: &[&Distribution],
    ) -> Result<Option<Vec<Distribution>>> {
        let children_ranges = children
            .iter()
            .map(|c| c.range())
            .collect::<Result<Vec<_>>>()?;
        let children_ranges_refs = children_ranges.iter().collect::<Vec<_>>();
        let parent_range = parent.range()?;
        let Some(propagated_children) =
            self.propagate_constraints(&parent_range, children_ranges_refs.as_slice())?
        else {
            return Ok(None);
        };
        izip!(propagated_children.into_iter(), children_ranges, children)
            .map(|(new_interval, old_interval, child)| {
                if new_interval == old_interval {
                    // We weren't able to narrow the range, preserve the old statistics.
                    Ok((*child).clone())
                } else if new_interval.data_type().eq(&DataType::Boolean) {
                    let dt = old_interval.data_type();
                    let p = if new_interval.eq(&Interval::CERTAINLY_TRUE) {
                        ScalarValue::new_one(&dt)
                    } else if new_interval.eq(&Interval::CERTAINLY_FALSE) {
                        ScalarValue::new_zero(&dt)
                    } else {
                        unreachable!("Given that we have a range reduction for a boolean interval, we should have certainty")
                    }?;
                    Distribution::new_bernoulli(p)
                } else {
                    Distribution::new_from_interval(new_interval)
                }
            })
            .collect::<Result<_>>()
            .map(Some)
    }

    /// Calculates the properties of this [`PhysicalExpr`] based on its
    /// children's properties (i.e. order and range), recursively aggregating
    /// the information from its children. In cases where the [`PhysicalExpr`]
    /// has no children (e.g., `Literal` or `Column`), these properties should
    /// be specified externally, as the function defaults to unknown properties.
    fn get_properties(&self, _children: &[ExprProperties]) -> Result<ExprProperties> {
        Ok(ExprProperties::new_unknown())
    }

    /// Format this `PhysicalExpr` in nice human readable "SQL" format
    ///
    /// Specifically, this format is designed to be readable by humans, at the
    /// expense of details. Use `Display` or `Debug` for more detailed
    /// representation.
    ///
    /// See the [`fmt_sql`] function for an example of printing `PhysicalExpr`s as SQL.
    ///
    fn fmt_sql(&self, f: &mut Formatter<'_>) -> fmt::Result;

    /// Take a snapshot of this `PhysicalExpr`, if it is dynamic.
    ///
    /// "Dynamic" in this case means containing references to structures that may change
    /// during plan execution, such as hash tables.
    ///
    /// This method is used to capture the current state of `PhysicalExpr`s that may contain
    /// dynamic references to other operators in order to serialize it over the wire
    /// or treat it via downcast matching.
    ///
    /// You should not call this method directly as it does not handle recursion.
    /// Instead use [`snapshot_physical_expr`] to handle recursion and capture the
    /// full state of the `PhysicalExpr`.
    ///
    /// This is expected to return "simple" expressions that do not have mutable state
    /// and are composed of DataFusion's built-in `PhysicalExpr` implementations.
    /// Callers however should *not* assume anything about the returned expressions
    /// since callers and implementers may not agree on what "simple" or "built-in"
    /// means.
    /// In other words, if you need to serialize a `PhysicalExpr` across the wire
    /// you should call this method and then try to serialize the result,
    /// but you should handle unknown or unexpected `PhysicalExpr` implementations gracefully
    /// just as if you had not called this method at all.
    ///
    /// In particular, consider:
    /// * A `PhysicalExpr` that references the current state of a `datafusion::physical_plan::TopK`
    ///   that is involved in a query with `SELECT * FROM t1 ORDER BY a LIMIT 10`.
    ///   This function may return something like `a >= 12`.
    /// * A `PhysicalExpr` that references the current state of a `datafusion::physical_plan::joins::HashJoinExec`
    ///   from a query such as `SELECT * FROM t1 JOIN t2 ON t1.a = t2.b`.
    ///   This function may return something like `t2.b IN (1, 5, 7)`.
    ///
    /// A system or function that can only deal with a hardcoded set of `PhysicalExpr` implementations
    /// or needs to serialize this state to bytes may not be able to handle these dynamic references.
    /// In such cases, we should return a simplified version of the `PhysicalExpr` that does not
    /// contain these dynamic references.
    ///
    /// Systems that implement remote execution of plans, e.g. serialize a portion of the query plan
    /// and send it across the wire to a remote executor may want to call this method after
    /// every batch on the source side and brodcast / update the current snaphot to the remote executor.
    ///
    /// Note for implementers: this method should *not* handle recursion.
    /// Recursion is handled in [`snapshot_physical_expr`].
    fn snapshot(&self) -> Result<Option<Arc<dyn PhysicalExpr>>> {
        // By default, we return None to indicate that this PhysicalExpr does not
        // have any dynamic references or state.
        // This is a safe default behavior.
        Ok(None)
    }
}

/// [`PhysicalExpr`] can't be constrained by [`Eq`] directly because it must remain object
/// safe. To ease implementation, blanket implementation is provided for [`Eq`] types.
pub trait DynEq {
    fn dyn_eq(&self, other: &dyn Any) -> bool;
}

impl<T: Eq + Any> DynEq for T {
    fn dyn_eq(&self, other: &dyn Any) -> bool {
        other.downcast_ref::<Self>() == Some(self)
    }
}

impl PartialEq for dyn PhysicalExpr {
    fn eq(&self, other: &Self) -> bool {
        self.dyn_eq(other.as_any())
    }
}

impl Eq for dyn PhysicalExpr {}

/// [`PhysicalExpr`] can't be constrained by [`Hash`] directly because it must remain
/// object safe. To ease implementation blanket implementation is provided for [`Hash`]
/// types.
pub trait DynHash {
    fn dyn_hash(&self, _state: &mut dyn Hasher);
}

impl<T: Hash + Any> DynHash for T {
    fn dyn_hash(&self, mut state: &mut dyn Hasher) {
        self.type_id().hash(&mut state);
        self.hash(&mut state)
    }
}

impl Hash for dyn PhysicalExpr {
    fn hash<H: Hasher>(&self, state: &mut H) {
        self.dyn_hash(state);
    }
}

/// Returns a copy of this expr if we change any child according to the pointer comparison.
/// The size of `children` must be equal to the size of `PhysicalExpr::children()`.
pub fn with_new_children_if_necessary(
    expr: Arc<dyn PhysicalExpr>,
    children: Vec<Arc<dyn PhysicalExpr>>,
) -> Result<Arc<dyn PhysicalExpr>> {
    let old_children = expr.children();
    if children.len() != old_children.len() {
        internal_err!("PhysicalExpr: Wrong number of children")
    } else if children.is_empty()
        || children
            .iter()
            .zip(old_children.iter())
            .any(|(c1, c2)| !Arc::ptr_eq(c1, c2))
    {
        Ok(expr.with_new_children(children)?)
    } else {
        Ok(expr)
    }
}

#[deprecated(since = "44.0.0")]
pub fn down_cast_any_ref(any: &dyn Any) -> &dyn Any {
    if any.is::<Arc<dyn PhysicalExpr>>() {
        any.downcast_ref::<Arc<dyn PhysicalExpr>>()
            .unwrap()
            .as_any()
    } else if any.is::<Box<dyn PhysicalExpr>>() {
        any.downcast_ref::<Box<dyn PhysicalExpr>>()
            .unwrap()
            .as_any()
    } else {
        any
    }
}

/// Returns [`Display`] able a list of [`PhysicalExpr`]
///
/// Example output: `[a + 1, b]`
pub fn format_physical_expr_list<T>(exprs: T) -> impl Display
where
    T: IntoIterator,
    T::Item: Display,
    T::IntoIter: Clone,
{
    struct DisplayWrapper<I>(I)
    where
        I: Iterator + Clone,
        I::Item: Display;

    impl<I> Display for DisplayWrapper<I>
    where
        I: Iterator + Clone,
        I::Item: Display,
    {
        fn fmt(&self, f: &mut Formatter<'_>) -> fmt::Result {
            let mut iter = self.0.clone();
            write!(f, "[")?;
            if let Some(expr) = iter.next() {
                write!(f, "{}", expr)?;
            }
            for expr in iter {
                write!(f, ", {}", expr)?;
            }
            write!(f, "]")?;
            Ok(())
        }
    }

    DisplayWrapper(exprs.into_iter())
}

/// Prints a [`PhysicalExpr`] in a SQL-like format
///
/// # Example
/// ```
/// # // The boiler plate needed to create a `PhysicalExpr` for the example
/// # use std::any::Any;
/// # use std::fmt::Formatter;
/// # use std::sync::Arc;
/// # use arrow::array::RecordBatch;
/// # use arrow::datatypes::{DataType, Schema};
/// # use datafusion_common::Result;
/// # use datafusion_expr_common::columnar_value::ColumnarValue;
/// # use datafusion_physical_expr_common::physical_expr::{fmt_sql, DynEq, PhysicalExpr};
/// # #[derive(Debug, Hash, PartialOrd, PartialEq)]
/// # struct MyExpr {};
/// # impl PhysicalExpr for MyExpr {fn as_any(&self) -> &dyn Any { unimplemented!() }
/// # fn data_type(&self, input_schema: &Schema) -> Result<DataType> { unimplemented!() }
/// # fn nullable(&self, input_schema: &Schema) -> Result<bool> { unimplemented!() }
/// # fn evaluate(&self, batch: &RecordBatch) -> Result<ColumnarValue> { unimplemented!() }
/// # fn children(&self) -> Vec<&Arc<dyn PhysicalExpr>>{ unimplemented!() }
/// # fn with_new_children(self: Arc<Self>, children: Vec<Arc<dyn PhysicalExpr>>) -> Result<Arc<dyn PhysicalExpr>> { unimplemented!() }
/// # fn fmt_sql(&self, f: &mut Formatter<'_>) -> std::fmt::Result { write!(f, "CASE a > b THEN 1 ELSE 0 END") }
/// # }
/// # impl std::fmt::Display for MyExpr {fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result { unimplemented!() } }
/// # impl DynEq for MyExpr {fn dyn_eq(&self, other: &dyn Any) -> bool { unimplemented!() } }
/// # fn make_physical_expr() -> Arc<dyn PhysicalExpr> { Arc::new(MyExpr{}) }
/// let expr: Arc<dyn PhysicalExpr> = make_physical_expr();
/// // wrap the expression in `sql_fmt` which can be used with
/// // `format!`, `to_string()`, etc
/// let expr_as_sql = fmt_sql(expr.as_ref());
/// assert_eq!(
///   "The SQL: CASE a > b THEN 1 ELSE 0 END",
///   format!("The SQL: {expr_as_sql}")
/// );
/// ```
pub fn fmt_sql(expr: &dyn PhysicalExpr) -> impl Display + '_ {
    struct Wrapper<'a> {
        expr: &'a dyn PhysicalExpr,
    }

    impl Display for Wrapper<'_> {
        fn fmt(&self, f: &mut Formatter) -> fmt::Result {
            self.expr.fmt_sql(f)?;
            Ok(())
        }
    }

    Wrapper { expr }
}

/// Take a snapshot of the given `PhysicalExpr` if it is dynamic.
///
/// Take a snapshot of this `PhysicalExpr` if it is dynamic.
/// This is used to capture the current state of `PhysicalExpr`s that may contain
/// dynamic references to other operators in order to serialize it over the wire
/// or treat it via downcast matching.
///
/// See the documentation of [`PhysicalExpr::snapshot`] for more details.
///
/// # Returns
///
/// Returns an `Option<Arc<dyn PhysicalExpr>>` which is the snapshot of the
/// `PhysicalExpr` if it is dynamic. If the `PhysicalExpr` does not have
/// any dynamic references or state, it returns `None`.
pub fn snapshot_physical_expr(
    expr: Arc<dyn PhysicalExpr>,
) -> Result<Arc<dyn PhysicalExpr>> {
    expr.transform_up(|e| {
        if let Some(snapshot) = e.snapshot()? {
            Ok(Transformed::yes(snapshot))
        } else {
            Ok(Transformed::no(Arc::clone(&e)))
        }
    })
    .data()
}