findepi commented on code in PR #17872:
URL: https://github.com/apache/datafusion/pull/17872#discussion_r2399984085
##########
datafusion/expr/src/async_udf.rs:
##########
@@ -62,17 +61,14 @@ pub struct AsyncScalarUDF {
impl PartialEq for AsyncScalarUDF {
fn eq(&self, other: &Self) -> bool {
- let Self { inner } = self;
- // TODO when MSRV >= 1.86.0, switch to
`inner.equals(other.inner.as_ref())` leveraging trait upcasting.
- arc_ptr_eq(inner, &other.inner)
+ self.inner.dyn_eq(other.inner.as_any())
}
}
impl Eq for AsyncScalarUDF {}
impl Hash for AsyncScalarUDF {
fn hash<H: Hasher>(&self, state: &mut H) {
- let Self { inner } = self;
Review Comment:
as above -- Keep this line please.
##########
datafusion/expr/src/async_udf.rs:
##########
@@ -62,17 +61,14 @@ pub struct AsyncScalarUDF {
impl PartialEq for AsyncScalarUDF {
fn eq(&self, other: &Self) -> bool {
- let Self { inner } = self;
Review Comment:
Keep this line please.
This ensures that the eq impl is updated whenever a field is added to the
struct
##########
datafusion/expr/src/async_udf.rs:
##########
@@ -132,3 +128,128 @@ impl Display for AsyncScalarUDF {
write!(f, "AsyncScalarUDF: {}", self.inner.name())
}
}
+
+#[cfg(test)]
+mod tests {
+ use std::{collections::HashSet, sync::Arc};
+
+ use arrow::datatypes::DataType;
+ use async_trait::async_trait;
+ use datafusion_common::error::Result;
+ use datafusion_expr_common::{columnar_value::ColumnarValue,
signature::Signature};
+
+ use crate::{
+ async_udf::{AsyncScalarUDF, AsyncScalarUDFImpl},
+ ScalarFunctionArgs, ScalarUDFImpl,
+ };
+
+ #[derive(Debug, PartialEq, Eq, Hash, Clone)]
+ struct TestAsyncUDFImpl1 {
+ a: i32,
+ }
+
+ impl ScalarUDFImpl for TestAsyncUDFImpl1 {
+ fn as_any(&self) -> &dyn std::any::Any {
+ self
+ }
+
+ fn name(&self) -> &str {
+ todo!()
+ }
+
+ fn signature(&self) -> &Signature {
+ todo!()
+ }
+
+ fn return_type(&self, _arg_types: &[DataType]) -> Result<DataType> {
+ todo!()
+ }
+
+ fn invoke_with_args(&self, _args: ScalarFunctionArgs) ->
Result<ColumnarValue> {
+ todo!()
+ }
+ }
+
+ #[async_trait]
+ impl AsyncScalarUDFImpl for TestAsyncUDFImpl1 {
+ async fn invoke_async_with_args(
+ &self,
+ _args: ScalarFunctionArgs,
+ ) -> Result<ColumnarValue> {
+ todo!()
+ }
+ }
+
+ #[derive(Debug, PartialEq, Eq, Hash, Clone)]
+ struct TestAsyncUDFImpl2 {
+ a: i32,
+ }
+
+ impl ScalarUDFImpl for TestAsyncUDFImpl2 {
+ fn as_any(&self) -> &dyn std::any::Any {
+ self
+ }
+
+ fn name(&self) -> &str {
+ todo!()
+ }
+
+ fn signature(&self) -> &Signature {
+ todo!()
+ }
+
+ fn return_type(&self, _arg_types: &[DataType]) -> Result<DataType> {
+ todo!()
+ }
+
+ fn invoke_with_args(&self, _args: ScalarFunctionArgs) ->
Result<ColumnarValue> {
+ todo!()
+ }
+ }
+
+ #[async_trait]
+ impl AsyncScalarUDFImpl for TestAsyncUDFImpl2 {
+ async fn invoke_async_with_args(
+ &self,
+ _args: ScalarFunctionArgs,
+ ) -> Result<ColumnarValue> {
+ todo!()
+ }
+ }
+
+ #[test]
+ fn udf_equality_and_hash() {
Review Comment:
it's generally beneficial to test eq, hash and ord together, because their
contracts are closely related
see existing `test_partial_eq_hash_and_partial_ord` for example
##########
datafusion/expr/src/async_udf.rs:
##########
@@ -132,3 +128,128 @@ impl Display for AsyncScalarUDF {
write!(f, "AsyncScalarUDF: {}", self.inner.name())
}
}
+
+#[cfg(test)]
+mod tests {
+ use std::{collections::HashSet, sync::Arc};
+
+ use arrow::datatypes::DataType;
+ use async_trait::async_trait;
+ use datafusion_common::error::Result;
+ use datafusion_expr_common::{columnar_value::ColumnarValue,
signature::Signature};
+
+ use crate::{
+ async_udf::{AsyncScalarUDF, AsyncScalarUDFImpl},
+ ScalarFunctionArgs, ScalarUDFImpl,
+ };
+
+ #[derive(Debug, PartialEq, Eq, Hash, Clone)]
+ struct TestAsyncUDFImpl1 {
+ a: i32,
+ }
+
+ impl ScalarUDFImpl for TestAsyncUDFImpl1 {
+ fn as_any(&self) -> &dyn std::any::Any {
+ self
+ }
+
+ fn name(&self) -> &str {
+ todo!()
+ }
+
+ fn signature(&self) -> &Signature {
+ todo!()
+ }
+
+ fn return_type(&self, _arg_types: &[DataType]) -> Result<DataType> {
+ todo!()
+ }
+
+ fn invoke_with_args(&self, _args: ScalarFunctionArgs) ->
Result<ColumnarValue> {
+ todo!()
+ }
+ }
+
+ #[async_trait]
+ impl AsyncScalarUDFImpl for TestAsyncUDFImpl1 {
+ async fn invoke_async_with_args(
+ &self,
+ _args: ScalarFunctionArgs,
+ ) -> Result<ColumnarValue> {
+ todo!()
+ }
+ }
+
+ #[derive(Debug, PartialEq, Eq, Hash, Clone)]
+ struct TestAsyncUDFImpl2 {
+ a: i32,
+ }
+
+ impl ScalarUDFImpl for TestAsyncUDFImpl2 {
+ fn as_any(&self) -> &dyn std::any::Any {
+ self
+ }
+
+ fn name(&self) -> &str {
+ todo!()
+ }
+
+ fn signature(&self) -> &Signature {
+ todo!()
+ }
+
+ fn return_type(&self, _arg_types: &[DataType]) -> Result<DataType> {
+ todo!()
+ }
+
+ fn invoke_with_args(&self, _args: ScalarFunctionArgs) ->
Result<ColumnarValue> {
+ todo!()
+ }
+ }
+
+ #[async_trait]
+ impl AsyncScalarUDFImpl for TestAsyncUDFImpl2 {
+ async fn invoke_async_with_args(
+ &self,
+ _args: ScalarFunctionArgs,
+ ) -> Result<ColumnarValue> {
+ todo!()
+ }
+ }
+
+ #[test]
+ fn udf_equality_and_hash() {
+ // Inner is same cloned arc -> equal
+ let inner = Arc::new(TestAsyncUDFImpl1 { a: 1 });
+ let a = AsyncScalarUDF::new(Arc::clone(&inner) as Arc<dyn
AsyncScalarUDFImpl>);
+ let b = AsyncScalarUDF::new(inner);
+ assert_eq!(a, b);
+ let mut set = HashSet::new();
+ set.insert(a);
+ assert!(set.contains(&b));
Review Comment:
Does this test a HashSet, or implementation of hash/eq?
If the later, is the property testable without using a hashset as an
intermediary?
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
