[FLINK-16070] Blink planner can not extract correct unique key for UpsertStreamTableSink - ASF JIRA

XML

Word

Printable

JSON

Details

Type: Bug
Status: Resolved
Priority: Critical
Resolution: Fixed
Affects Version/s: 1.10.0
Fix Version/s: 1.10.1, 1.11.0
Component/s: Table SQL / Planner
Labels:
- pull-request-available

Description

I reproduce an Elasticsearch6UpsertTableSink issue which user reported in mail list[1] that Blink planner can not extract correct unique key for following query, but legacy planner works well.

// user code
INSERT INTO ES6_ZHANGLE_OUTPUT  
 SELECT aggId, pageId, ts_min as ts,  
       count(case when eventId = 'exposure' then 1 else null end) as expoCnt,  
       count(case when eventId = 'click' then 1 else null end) as clkCnt  
 FROM  (    
     SELECT        
       'ZL_001' as aggId,
        pageId,        
        eventId,        
        recvTime,        
        ts2Date(recvTime) as ts_min    
     from kafka_zl_etrack_event_stream    
     where eventId in ('exposure', 'click')  
 ) as t1  
 group by aggId, pageId, ts_min

I found that blink planner can not extract correct unique key in `FlinkRelMetadataQuery.getUniqueKeys(relNode)`, legacy planner works well in `org.apache.flink.table.plan.util.UpdatingPlanChecker.getUniqueKeyFields(...) `. A simple ETL job to reproduce this issue can refers[2]

[1]http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Flink-1-10-es-sink-exception-td32773.html

[2]https://github.com/leonardBang/flink-sql-etl/blob/master/etl-job/src/main/java/kafka2es/Kafka2UpsertEs.java