Github user koertkuipers commented on a diff in the pull request:
https://github.com/apache/spark/pull/14576#discussion_r74316735
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/expressions/ReduceAggregator.scala
---
@@ -0,0 +1,79 @@
+/*
+ * 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.
+ */
+
+package org.apache.spark.sql.expressions
+
+import org.apache.spark.annotation.Experimental
+import org.apache.spark.sql.Encoder
+import org.apache.spark.sql.catalyst.encoders.ExpressionEncoder
+
+/**
+ * :: Experimental ::
+ * An aggregator that uses a single associative and commutative reduce
function. This reduce
+ * function can be used to go through all input values and reduces them to
a single value.
+ * If there is no input, a null value is returned.
+ *
+ * @since 2.1.0
+ */
+@Experimental
+abstract class ReduceAggregator[T] extends Aggregator[T, (Boolean, T), T] {
+
+ // Question 1: Should func and encoder be parameters rather than
abstract methods?
+ // rxin: abstract method has better java compatibility and forces
naming the concrete impl,
+ // whereas parameter has better type inference (infer encoders via
context bounds).
+ // Question 2: Should finish throw an exception or return null if there
is no input?
+ // rxin: null might be more "SQL" like, whereas exception is more Scala
like.
--- End diff --
i prefer null over an exception
ideally this would return an option. the operation really should be:
```def reduceGroups(f: (V, V) => V): Dataset[(K, Option[V])]```
and the class:
```abstract class ReduceAggregator[T] extends Aggregator[T, (Boolean, T),
Option[T]]```
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
