[
https://issues.apache.org/jira/browse/FLINK-29068?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17600663#comment-17600663
]
Qingsheng Ren commented on FLINK-29068:
---------------------------------------
CREATE FUNCTION USING JAR:
# Prepared a JAR with UDF implementation
# Verified creating an UDF with JAR in local filesystem
# Verified `SHOW JARS`
# Verified executing query with the function created above
ADD JAR:
# Using Kafka connector JAR for validation, including:
## Kafka SQL connector uber JAR
## Kafka connector JAR + Kafka clients JAR
# Verified `ADD JAR`
# Verified creating Kafka source and sink table
# Verified executing query with Kafka source to Kafka sink, including using
UDTF created by `USING JAR` syntax and custom sink partitioner added by `ADD
JAR` syntax.
# Verified using multiple UDFs in the query
> Release Testing: Verify FLIP-214 Create function using jar syntax
> -----------------------------------------------------------------
>
> Key: FLINK-29068
> URL: https://issues.apache.org/jira/browse/FLINK-29068
> Project: Flink
> Issue Type: Sub-task
> Components: Table SQL / API
> Affects Versions: 1.16.0
> Reporter: dalongliu
> Assignee: Qingsheng Ren
> Priority: Blocker
> Labels: release-testing
> Fix For: 1.16.0
>
>
> This issue aims to verify FLIP-214:
> [https://cwiki.apache.org/confluence/display/FLINK/FLIP-214+Support+Advanced+Function+DDL]
> We can verify it in SQL client after we build the flink-dist package.
> *verify create function using jar syntax*
> # Preparing a jar that contains the udf class implementation, the jar path
> should be a local or remote path such as hdfs/s3/oss.
> # Creating a catalog function/temporary catalog function/temporary system
> function using the jar provided jar path.
> # Using `show jars` to verify whether the function has been created
> successfully.
> # Write a query which refer to the above udf, and then execute it. Verify
> the query execution and result correctness.
>
> *verify add jar syntax*
> We have ported the implementation of add jar/show jars from SQL client to
> TableEnvironment in FLIP-214, and the customed jar class load using user
> classloader instead of thread context classloader now, so we also should
> verify the add jar syntax.
> # Preparing a jar that contains the udf or catalog implementation, the jar
> path should be a local or remote path such as hdfs/s3/oss.
> # Using add jar syntax by the provided jar resource.
> # Creating a catalog(such as jdbc/hive) using the added jar and execute
> other ddls within the catalog
> # Creating a udf using the added jar and write a query referring to the udf,
> then submit it.
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
