Hi, guys
When running the same Flink sql like the following, I met exception like
"org.apache.flink.table.api.TableException: UpsertStreamTableSink requires that
Table has a full primary keys if it is updated". I am using the latest Flink
1.10 release with blink planner enabled. Because the same logic runs well
within Flink 1.8.2 old planner. Does the SQL usage has some problem or may has
a bug here ?
INSERT INTO ES6_ZHANGLE_OUTPUT(aggId, pageId, ts, expoCnt, clkCnt)
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 clickCnt
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
