[
https://issues.apache.org/jira/browse/FLINK-18701?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Timo Walther updated FLINK-18701:
---------------------------------
Comment: was deleted
(was: This issue was marked "stale-assigned" and has not received an update in
7 days. It is now automatically unassigned. If you are still working on it, you
can assign it to yourself again. Please also give an update about the status of
the work.)
> NOT NULL constraint is not guaranteed when aggregation split is enabled
> -----------------------------------------------------------------------
>
> Key: FLINK-18701
> URL: https://issues.apache.org/jira/browse/FLINK-18701
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / Planner
> Affects Versions: 1.11.1
> Reporter: Jark Wu
> Priority: Minor
>
> Take the following test:
> {{org.apache.flink.table.planner.runtime.stream.sql.SplitAggregateITCase#testMinMaxWithRetraction}}
> {code:scala}
> val t1 = tEnv.sqlQuery(
> s"""
> |SELECT
> | c, MIN(b), MAX(b), COUNT(DISTINCT a)
> |FROM(
> | SELECT
> | a, COUNT(DISTINCT b) as b, MAX(b) as c
> | FROM T
> | GROUP BY a
> |) GROUP BY c
> """.stripMargin)
> val sink = new TestingRetractSink
> t1.toRetractStream[Row].addSink(sink)
> env.execute()
> println(sink.getRawResults)
> {code}
> The query schema is
> {code:java}
> root
> |-- c: INT
> |-- EXPR$1: BIGINT NOT NULL
> |-- EXPR$2: BIGINT NOT NULL
> |-- EXPR$3: BIGINT NOT NULL
> {code}
> This should be correct as the count is never null and thus min/max are never
> null, however, we can receive null in the sink.
> {code}
> List((true,1,null,null,1), (true,2,2,2,1), (false,1,null,null,1),
> (true,6,2,2,1), (true,5,1,1,0), (false,5,1,1,0), (true,5,1,1,2),
> (true,4,2,2,0), (false,5,1,1,2), (true,5,1,3,2), (false,4,2,2,0),
> (false,5,1,3,2), (true,5,1,4,2))
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
