Convert stddev and stddev_pop to UDAF (apache#10834)

* add stddev and stddev_pot udaf

* remove aggregation function stddev and stddev_pop

* register func and modified return type

* cargo fmt

* regen proto

* cargo clippy

* fix window function support

* cargo fmt

* throw not_impl_err instead

* use default sliding accumulator

datafusion/core/src/dataframe/mod.rs

@@ -50,12 +50,12 @@ use datafusion_common::{
 };
 use datafusion_expr::lit;
 use datafusion_expr::{
-    avg, count, max, min, stddev, utils::COUNT_STAR_EXPANSION,
-    TableProviderFilterPushDown, UNNAMED_TABLE,
+    avg, count, max, min, utils::COUNT_STAR_EXPANSION, TableProviderFilterPushDown,
+    UNNAMED_TABLE,
 };
 use datafusion_expr::{case, is_null};
-use datafusion_functions_aggregate::expr_fn::median;
 use datafusion_functions_aggregate::expr_fn::sum;
+use datafusion_functions_aggregate::expr_fn::{median, stddev};
 
 use async_trait::async_trait;
 

datafusion/expr/src/aggregate_function.rs

@@ -49,10 +49,6 @@ pub enum AggregateFunction {
     NthValue,
     /// Variance (Population)
     VariancePop,
-    /// Standard Deviation (Sample)
-    Stddev,
-    /// Standard Deviation (Population)
-    StddevPop,
     /// Correlation
     Correlation,
     /// Slope from linear regression
@@ -107,8 +103,6 @@ impl AggregateFunction {
             ArrayAgg => "ARRAY_AGG",
             NthValue => "NTH_VALUE",
             VariancePop => "VAR_POP",
-            Stddev => "STDDEV",
-            StddevPop => "STDDEV_POP",
             Correlation => "CORR",
             RegrSlope => "REGR_SLOPE",
             RegrIntercept => "REGR_INTERCEPT",
@@ -159,9 +153,6 @@ impl FromStr for AggregateFunction {
             "string_agg" => AggregateFunction::StringAgg,
             // statistical
             "corr" => AggregateFunction::Correlation,
-            "stddev" => AggregateFunction::Stddev,
-            "stddev_pop" => AggregateFunction::StddevPop,
-            "stddev_samp" => AggregateFunction::Stddev,
             "var_pop" => AggregateFunction::VariancePop,
             "regr_slope" => AggregateFunction::RegrSlope,
             "regr_intercept" => AggregateFunction::RegrIntercept,
@@ -231,8 +222,6 @@ impl AggregateFunction {
             AggregateFunction::Correlation => {
                 correlation_return_type(&coerced_data_types[0])
             }
-            AggregateFunction::Stddev => stddev_return_type(&coerced_data_types[0]),
-            AggregateFunction::StddevPop => stddev_return_type(&coerced_data_types[0]),
             AggregateFunction::RegrSlope
             | AggregateFunction::RegrIntercept
             | AggregateFunction::RegrCount
@@ -304,8 +293,6 @@ impl AggregateFunction {
             }
             AggregateFunction::Avg
             | AggregateFunction::VariancePop
-            | AggregateFunction::Stddev
-            | AggregateFunction::StddevPop
             | AggregateFunction::ApproxMedian => {
                 Signature::uniform(1, NUMERICS.to_vec(), Volatility::Immutable)
             }

datafusion/expr/src/expr_fn.rs

@@ -383,18 +383,6 @@ pub fn scalar_subquery(subquery: Arc<LogicalPlan>) -> Expr {
     })
 }
 
-/// Create an expression to represent the stddev() aggregate function
-pub fn stddev(expr: Expr) -> Expr {
-    Expr::AggregateFunction(AggregateFunction::new(
-        aggregate_function::AggregateFunction::Stddev,
-        vec![expr],
-        false,
-        None,
-        None,
-        None,
-    ))
-}
-
 /// Create a grouping set
 pub fn grouping_set(exprs: Vec<Vec<Expr>>) -> Expr {
     Expr::GroupingSet(GroupingSet::GroupingSets(exprs))

datafusion/expr/src/type_coercion/aggregates.rs

@@ -161,16 +161,6 @@ pub fn coerce_types(
             }
             Ok(vec![Float64, Float64])
         }
-        AggregateFunction::Stddev | AggregateFunction::StddevPop => {
-            if !is_stddev_support_arg_type(&input_types[0]) {
-                return plan_err!(
-                    "The function {:?} does not support inputs of type {:?}.",
-                    agg_fun,
-                    input_types[0]
-                );
-            }
-            Ok(vec![Float64])
-        }
         AggregateFunction::Correlation => {
             if !is_correlation_support_arg_type(&input_types[0]) {
                 return plan_err!(
@@ -408,15 +398,6 @@ pub fn correlation_return_type(arg_type: &DataType) -> Result<DataType> {
     }
 }
 
-/// function return type of standard deviation
-pub fn stddev_return_type(arg_type: &DataType) -> Result<DataType> {
-    if NUMERICS.contains(arg_type) {
-        Ok(DataType::Float64)
-    } else {
-        plan_err!("STDDEV does not support {arg_type:?}")
-    }
-}
-
 /// function return type of an average
 pub fn avg_return_type(arg_type: &DataType) -> Result<DataType> {
     match arg_type {
@@ -511,13 +492,6 @@ pub fn is_covariance_support_arg_type(arg_type: &DataType) -> bool {
     )
 }
 
-pub fn is_stddev_support_arg_type(arg_type: &DataType) -> bool {
-    matches!(
-        arg_type,
-        arg_type if NUMERICS.contains(arg_type)
-    )
-}
-
 pub fn is_correlation_support_arg_type(arg_type: &DataType) -> bool {
     matches!(
         arg_type,
@@ -664,17 +638,6 @@ mod tests {
         Ok(())
     }
 
-    #[test]
-    fn test_stddev_return_data_type() -> Result<()> {
-        let data_type = DataType::Float64;
-        let result_type = stddev_return_type(&data_type)?;
-        assert_eq!(DataType::Float64, result_type);
-
-        let data_type = DataType::Decimal128(36, 10);
-        assert!(stddev_return_type(&data_type).is_err());
-        Ok(())
-    }
-
     #[test]
     fn test_covariance_return_data_type() -> Result<()> {
         let data_type = DataType::Float64;

datafusion/functions-aggregate/src/lib.rs

@@ -58,6 +58,7 @@ pub mod macros;
 pub mod covariance;
 pub mod first_last;
 pub mod median;
+pub mod stddev;
 pub mod sum;
 pub mod variance;
 
@@ -74,6 +75,8 @@ pub mod expr_fn {
     pub use super::first_last::first_value;
     pub use super::first_last::last_value;
     pub use super::median::median;
+    pub use super::stddev::stddev;
+    pub use super::stddev::stddev_pop;
     pub use super::sum::sum;
     pub use super::variance::var_sample;
 }
@@ -88,6 +91,8 @@ pub fn all_default_aggregate_functions() -> Vec<Arc<AggregateUDF>> {
         covariance::covar_pop_udaf(),
         median::median_udaf(),
         variance::var_samp_udaf(),
+        stddev::stddev_udaf(),
+        stddev::stddev_pop_udaf(),
     ]
 }