hejian created KYLIN-3992: ----------------------------- Summary: easy exceeds 'max-return-row' when gouping by a high-cardinality dimension with a long day span filter Key: KYLIN-3992 URL: https://issues.apache.org/jira/browse/KYLIN-3992 Project: Kylin Issue Type: Bug Reporter: hejian
In my case, a dimension with 3million cardinality called seller_id, {code:java} select seller_id, sum(volume) as count_num from tab_orders where day_id>='2019-04-07' and day_id<'2019-04-27' group by seller_id order by count_num desc offset 0 limit 10{code} I think it relates to both the *cardinality(seller_id)* and *cardinality(day_id)* ? Hard to determine the suitable value to 'kylin.query.max-return-row', any good solution to handle this issue? -- This message was sent by Atlassian JIRA (v7.6.3#76005)