Github user rtudoran commented on a diff in the pull request:
https://github.com/apache/flink/pull/3550#discussion_r107184664
--- Diff:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/runtime/aggregate/DataStreamProcTimeAggregateGlobalWindowFunction.scala
---
@@ -0,0 +1,106 @@
+/*
+ * 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.flink.table.runtime.aggregate
+
+import
org.apache.flink.streaming.api.functions.windowing.RichAllWindowFunction
+import org.apache.flink.types.Row
+import org.apache.flink.util.Collector
+import org.apache.flink.streaming.api.windowing.windows.Window
+import org.apache.flink.configuration.Configuration
+import org.apache.flink.table.functions.Accumulator
+
+import java.lang.Iterable
+import org.apache.flink.table.functions.AggregateFunction
+
+/**
+ * Computes the final aggregate value from incrementally computed
aggreagtes.
+ *
+ * @param aggregates The aggregates to be computed
+ * @param aggFields the fields on which to apply the aggregate.
+ * @param forwardedFieldCount The fields to be carried from current row.
+ */
+class DataStreamProcTimeAggregateGlobalWindowFunction[W <: Window](
--- End diff --
@fhueske thanks for the design
I will implement it based on the processFunction. The implementation will
keep the events as you suggested. For the future implementations we will than
use this list state mechanism with the events to do the same operations as we
would do them in the table. The 2 are similar in functionality and have the
same worst case. Processfunction has a better best case scenario though. It
seems that the states and the buffers are similar in processing capabilities -
perhaps we should double check it with @aljoscha that there is not performance
difference between having a very large state with events and a very large
window buffer (e.g. million of events)
---
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.
---
