viirya commented on code in PR #234: URL: https://github.com/apache/arrow-datafusion-comet/pull/234#discussion_r1560032910
########## core/src/execution/datafusion/expressions/covariance.rs: ########## @@ -0,0 +1,385 @@ +/* + * 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, sync::Arc}; + +use arrow::{ + array::{ArrayRef, Float64Array}, + compute::cast, + datatypes::{DataType, Field}, +}; +use datafusion::logical_expr::Accumulator; +use datafusion_common::{ + downcast_value, unwrap_or_internal_err, DataFusionError, Result, ScalarValue, +}; +use datafusion_physical_expr::{ + aggregate::utils::down_cast_any_ref, + expressions::{format_state_name, StatsType}, + AggregateExpr, PhysicalExpr, +}; + +/// COVAR and COVAR_SAMP aggregate expression +#[derive(Debug, Clone)] +pub struct Covariance { + name: String, + expr1: Arc<dyn PhysicalExpr>, + expr2: Arc<dyn PhysicalExpr>, +} + +/// COVAR_POP aggregate expression +#[derive(Debug)] +pub struct CovariancePop { + name: String, + expr1: Arc<dyn PhysicalExpr>, + expr2: Arc<dyn PhysicalExpr>, +} + +impl Covariance { + /// Create a new COVAR aggregate function + pub fn new( + expr1: Arc<dyn PhysicalExpr>, + expr2: Arc<dyn PhysicalExpr>, + name: impl Into<String>, + data_type: DataType, + ) -> Self { + // the result of covariance just support FLOAT64 data type. + assert!(matches!(data_type, DataType::Float64)); + Self { + name: name.into(), + expr1, + expr2, + } + } +} + +impl AggregateExpr for Covariance { + /// Return a reference to Any that can be used for downcasting + fn as_any(&self) -> &dyn Any { + self + } + + fn field(&self) -> Result<Field> { + Ok(Field::new(&self.name, DataType::Float64, true)) + } + + fn create_accumulator(&self) -> Result<Box<dyn Accumulator>> { + Ok(Box::new(CovarianceAccumulator::try_new(StatsType::Sample)?)) + } + + fn state_fields(&self) -> Result<Vec<Field>> { + Ok(vec![ + Field::new( + format_state_name(&self.name, "count"), + DataType::Float64, + true, + ), + Field::new( + format_state_name(&self.name, "mean1"), + DataType::Float64, + true, + ), + Field::new( + format_state_name(&self.name, "mean2"), + DataType::Float64, + true, + ), + Field::new( + format_state_name(&self.name, "algo_const"), + DataType::Float64, + true, + ), + ]) + } + + fn expressions(&self) -> Vec<Arc<dyn PhysicalExpr>> { + vec![self.expr1.clone(), self.expr2.clone()] + } + + fn name(&self) -> &str { + &self.name + } +} + +impl PartialEq<dyn Any> for Covariance { + fn eq(&self, other: &dyn Any) -> bool { + down_cast_any_ref(other) + .downcast_ref::<Self>() + .map(|x| self.name == x.name && self.expr1.eq(&x.expr1) && self.expr2.eq(&x.expr2)) + .unwrap_or(false) + } +} + +impl CovariancePop { + /// Create a new COVAR_POP aggregate function + pub fn new( + expr1: Arc<dyn PhysicalExpr>, + expr2: Arc<dyn PhysicalExpr>, + name: impl Into<String>, + data_type: DataType, + ) -> Self { + // the result of covariance just support FLOAT64 data type. + assert!(matches!(data_type, DataType::Float64)); + Self { + name: name.into(), + expr1, + expr2, + } + } +} + +impl AggregateExpr for CovariancePop { + /// Return a reference to Any that can be used for downcasting + fn as_any(&self) -> &dyn Any { + self + } + + fn field(&self) -> Result<Field> { + Ok(Field::new(&self.name, DataType::Float64, true)) + } + + fn create_accumulator(&self) -> Result<Box<dyn Accumulator>> { + Ok(Box::new(CovarianceAccumulator::try_new( + StatsType::Population, + )?)) + } Review Comment: Yea, but the behavior is decided in `CovarianceAccumulator` based on its `StatsType`. I mean this two struct `CovariancePop` and `Covariance`. They are basically the same, except that `name` and `StatsType` are different. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
