It seems that I have met a more serious problem the same as you. So here two things I really want to know. 1. Have you found out how your query error happens? and second, is the cube you use the same as generated by sample.sh?
2. When I use SSB data, I build a Cube with d_year grouped, but when I execute the following queries. SELECT c_region FROM LINEORDER left JOIN customer on lo_custkey = c_custkey; without any result until my HBase shutdown(I haved tried twice)! The log throws found a badquery ! What confused me is that, is this an unescapable post progress? select c_region from lineorder left join customer on lo_custkey=c_custkey group by c_region; This sql return result in few seconds. Above all, I mean, If the kylin found what you want to query is realization, so it should give the answer first, no matter it use group by or not(I mean the judge should be more Intelligent, not execute a unbear full time basic query), and finally my cause the HBase shutdown. -- View this message in context: http://apache-kylin.74782.x6.nabble.com/jira-Created-KYLIN-2313-Cannot-find-a-cube-in-a-subquery-case-with-count-distinct-tp6751p6763.html Sent from the Apache Kylin mailing list archive at Nabble.com.
