[
https://issues.apache.org/jira/browse/FLINK-7206?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16134781#comment-16134781
]
ASF GitHub Bot commented on FLINK-7206:
---------------------------------------
Github user wuchong commented on a diff in the pull request:
https://github.com/apache/flink/pull/4355#discussion_r134152736
--- Diff:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/functions/utils/UserDefinedFunctionUtils.scala
---
@@ -307,6 +312,119 @@ object UserDefinedFunctionUtils {
//
----------------------------------------------------------------------------------------------
/**
+ * get data view type information from accumulator constructor.
+ *
+ * @param aggFun aggregate function
+ * @return the data view specification
+ */
+ def getDataViewTypeInfoFromConstructor(
+ aggFun: AggregateFunction[_, _])
+ : mutable.HashMap[String, TypeInformation[_]] = {
+
+ val resultMap = new mutable.HashMap[String, TypeInformation[_]]
+ val acc = aggFun.createAccumulator()
+ val fields: util.List[Field] =
TypeExtractor.getAllDeclaredFields(acc.getClass, true)
--- End diff --
Yes, I think we can use accumulator type info instead of field information.
> Implementation of DataView to support state access for UDAGG
> ------------------------------------------------------------
>
> Key: FLINK-7206
> URL: https://issues.apache.org/jira/browse/FLINK-7206
> Project: Flink
> Issue Type: Sub-task
> Components: Table API & SQL
> Reporter: Kaibo Zhou
> Assignee: Kaibo Zhou
>
> Implementation of MapView and ListView to support state access for UDAGG.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
