[
https://issues.apache.org/jira/browse/FLINK-27386?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17531995#comment-17531995
]
luoyuxia commented on FLINK-27386:
----------------------------------
To fix it is quite simple and I have submited a pr. But I would like to clarify
that although specify join hint using Hive dialect curerntly, but the join hint
wouldn't work before Flink planner supports it. Hopefully, the feature of join
hint in Flink itself is on the road and expected to be finished in Flink 1.16.
> throw NPE if multi MAPJOIN hint union all
> -----------------------------------------
>
> Key: FLINK-27386
> URL: https://issues.apache.org/jira/browse/FLINK-27386
> Project: Flink
> Issue Type: Sub-task
> Components: Connectors / Hive
> Affects Versions: 1.13.1, 1.15.0
> Reporter: tartarus
> Priority: Major
> Labels: pull-request-available
>
> We can reproduce through a UT
> Add test case in HiveDialectITCase
> {code:java}
> @Test
> public void testHiveMultiMapJoinUnionAll() {
> tableEnv.executeSql("create table t1 (id bigint, name string)");
> tableEnv.executeSql("create table t2 (id bigint, name string)");
> tableEnv.executeSql("select /*+ mapjoin(t2) */ t1.id from t1 join t2 on
> t1.id = t2.id union all select /*+ mapjoin(t2) */ t1.id from t1 join t2 on
> t1.name = t2.name");
> } {code}
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
