vaibhawvipul commented on code in PR #471: URL: https://github.com/apache/datafusion-comet/pull/471#discussion_r1621791279
########## core/src/execution/datafusion/expressions/negative.rs: ########## @@ -0,0 +1,272 @@ +// 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 crate::errors::CometError; +use arrow::compute::kernels::numeric::neg_wrapping; +use arrow_array::RecordBatch; +use arrow_schema::{DataType, Schema}; +use datafusion::{ + logical_expr::{interval_arithmetic::Interval, ColumnarValue}, + physical_expr::PhysicalExpr, +}; +use datafusion_common::{Result, ScalarValue}; +use datafusion_physical_expr::{ + aggregate::utils::down_cast_any_ref, sort_properties::SortProperties, +}; +use std::{ + any::Any, + hash::{Hash, Hasher}, + sync::Arc, +}; + +pub fn create_negate_expr( + expr: Arc<dyn PhysicalExpr>, + fail_on_error: bool, +) -> Result<Arc<dyn PhysicalExpr>, CometError> { + Ok(Arc::new(NegativeExpr::new(expr, fail_on_error))) +} + +/// Negative expression +#[derive(Debug, Hash)] +pub struct NegativeExpr { + /// Input expression + arg: Arc<dyn PhysicalExpr>, + fail_on_error: bool, +} + +fn arithmetic_overflow_error(from_type: &str) -> CometError { + CometError::ArithmeticOverflow { + from_type: from_type.to_string(), + } +} + +macro_rules! check_overflow { + ($array:expr, $array_type:ty, $min_val:expr, $type_name:expr) => {{ + let typed_array = $array + .as_any() + .downcast_ref::<$array_type>() + .expect(concat!(stringify!($array_type), " expected")); + for i in 0..typed_array.len() { + if typed_array.value(i) <= $min_val { + if $type_name == "byte" || $type_name == "short" { + let value = typed_array.value(i).to_string() + " caused"; + return Err(arithmetic_overflow_error(value.as_str()).into()); + } + return Err(arithmetic_overflow_error($type_name).into()); + } + } + }}; +} + +impl NegativeExpr { + /// Create new not expression + pub fn new(arg: Arc<dyn PhysicalExpr>, fail_on_error: bool) -> Self { + Self { arg, fail_on_error } + } + + /// Get the input expression + pub fn arg(&self) -> &Arc<dyn PhysicalExpr> { + &self.arg + } +} + +impl std::fmt::Display for NegativeExpr { + fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result { + write!(f, "(- {})", self.arg) + } +} + +impl PhysicalExpr for NegativeExpr { + /// Return a reference to Any that can be used for downcasting + fn as_any(&self) -> &dyn Any { + self + } + + fn data_type(&self, input_schema: &Schema) -> Result<DataType> { + self.arg.data_type(input_schema) + } + + fn nullable(&self, input_schema: &Schema) -> Result<bool> { + self.arg.nullable(input_schema) + } + + fn evaluate(&self, batch: &RecordBatch) -> Result<ColumnarValue> { + let arg = self.arg.evaluate(batch)?; + + // overflow checks only apply in ANSI mode + // datatypes supported are byte, short, integer, long, float, interval + match arg { + ColumnarValue::Array(array) => { + if self.fail_on_error { + match array.data_type() { + DataType::Int8 => { + check_overflow!(array, arrow::array::Int8Array, i8::MIN, "byte") + } + DataType::Int16 => { + check_overflow!(array, arrow::array::Int16Array, i16::MIN, "short") + } + DataType::Int32 => { + check_overflow!(array, arrow::array::Int32Array, i32::MIN, "integer") + } + DataType::Int64 => { + check_overflow!(array, arrow::array::Int64Array, i64::MIN, "long") + } + DataType::Float64 => { + check_overflow!(array, arrow::array::Float64Array, f64::MIN, "float") + } + DataType::Interval(value) => match value { + arrow::datatypes::IntervalUnit::YearMonth => check_overflow!( + array, + arrow::array::IntervalYearMonthArray, + i32::MIN, + "interval" + ), + arrow::datatypes::IntervalUnit::DayTime => check_overflow!( + array, + arrow::array::IntervalDayTimeArray, + i64::MIN, + "interval" + ), + arrow::datatypes::IntervalUnit::MonthDayNano => check_overflow!( + array, + arrow::array::IntervalMonthDayNanoArray, + i128::MIN, + "interval" + ), + }, + _ => { + // Overflow checks are not supported for other datatypes + } + } + } + let result = neg_wrapping(array.as_ref())?; + Ok(ColumnarValue::Array(result)) + } + ColumnarValue::Scalar(scalar) => { + if self.fail_on_error { + match scalar { + ScalarValue::Int8(value) => { + if value == Some(i8::MIN) { + return Err(arithmetic_overflow_error(" caused").into()); + } + } + ScalarValue::Int16(value) => { + if value == Some(i16::MIN) { + return Err(arithmetic_overflow_error(" caused").into()); + } + } + ScalarValue::Int32(value) => { + if value == Some(i32::MIN) { + return Err(arithmetic_overflow_error("integer").into()); + } + } + ScalarValue::Int64(value) => { + if value == Some(i64::MIN) { + return Err(arithmetic_overflow_error("long").into()); + } + } + ScalarValue::Float64(value) => { + if value == Some(f64::MIN) || value == Some(f64::MAX) { + return Err(arithmetic_overflow_error("float").into()); + } + } + ScalarValue::IntervalDayTime(value) => { + if value == Some(i64::MIN) { + return Err(arithmetic_overflow_error("interval").into()); + } + } + ScalarValue::IntervalYearMonth(value) => { + if value == Some(i32::MIN) { + return Err(arithmetic_overflow_error("interval").into()); + } + } + ScalarValue::IntervalMonthDayNano(value) => { + if value == Some(i128::MIN) || value == Some(i128::MAX) { + return Err(arithmetic_overflow_error("interval").into()); + } + } + _ => { + // Overflow checks are not supported for other datatypes + } + } + } + Ok(ColumnarValue::Scalar((scalar.arithmetic_negate())?)) + } + } + } + + fn children(&self) -> Vec<Arc<dyn PhysicalExpr>> { + vec![self.arg.clone()] + } + + fn with_new_children( + self: Arc<Self>, + children: Vec<Arc<dyn PhysicalExpr>>, + ) -> Result<Arc<dyn PhysicalExpr>> { + Ok(Arc::new(NegativeExpr::new( + children[0].clone(), + self.fail_on_error, + ))) + } + + fn dyn_hash(&self, state: &mut dyn Hasher) { + let mut s = state; + self.hash(&mut s); + } + + /// Given the child interval of a NegativeExpr, it calculates the NegativeExpr's interval. + /// It replaces the upper and lower bounds after multiplying them with -1. + /// Ex: `(a, b]` => `[-b, -a)` + fn evaluate_bounds(&self, children: &[&Interval]) -> Result<Interval> { + Interval::try_new( + children[0].upper().arithmetic_negate()?, + children[0].lower().arithmetic_negate()?, Review Comment: added checks, i think it will be in parity with what we are doing above in the code. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For additional commands, e-mail: github-h...@datafusion.apache.org
