[
https://issues.apache.org/jira/browse/FLINK-5795?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15873415#comment-15873415
]
ASF GitHub Bot commented on FLINK-5795:
---------------------------------------
Github user wuchong commented on a diff in the pull request:
https://github.com/apache/flink/pull/3330#discussion_r101907504
--- Diff:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/codegen/CodeGenerator.scala
---
@@ -1463,21 +1465,23 @@ class CodeGenerator(
*/
def addReusableFunction(function: UserDefinedFunction): String = {
val classQualifier = function.getClass.getCanonicalName
- val fieldTerm = s"function_${classQualifier.replace('.', '$')}"
+ val functionSerializedData = serialize(function)
+ val fieldTerm =
+ s"""
+ |function_${classQualifier.replace('.',
'$')}_${DigestUtils.md5Hex(functionSerializedData)}
--- End diff --
>But from tableAPI to SqlFunction the creation process has some work to do
because the current UDF parameters do not take effect.
Do you mean "from Java Table API to SqlFunction" ? If yes, the Java Table
API (i.e. parse expression from a string) will instantiate a new instance of
ScalarFunction when lookup this function call (the line is
https://github.com/apache/flink/blob/master/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/validate/FunctionCatalog.scala#L84).
TableFunction will reuse the object created by user. You can follow the way of
TableFunction to fix this.
BTW, scalar UDF in Scala Table API works good, right?
> Improve “UDTF" to support constructor with parameter.
> -----------------------------------------------------
>
> Key: FLINK-5795
> URL: https://issues.apache.org/jira/browse/FLINK-5795
> Project: Flink
> Issue Type: Sub-task
> Components: Table API & SQL
> Reporter: sunjincheng
> Assignee: sunjincheng
>
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
