I tried to create a simple star schema with 2 dimensions and 1 fact in Hive. Data in these tables is just few rows to test Kylin cube build process. Few issues are observed which are listed below:
Tables and data exists in Hive which are synched in Kylin project, so why Kylin creates Intermediate flat table in Hive and in builds dimensions again (in Step 1 and Step 3 respectively)? For very large tables won't this be a performance bottleneck? Also, the cube build fails repeatedly and faced issue with building the sample cube earlier as well. What are the pre-requisites to build a cube successfully in Kylin? Kylin log attached: kylin.log <http://apache-kylin-incubating.74782.x6.nabble.com/file/n2332/kylin.log> Note: I am using the correct Hadoop, Hive and HBase versions for this purpose. -- View this message in context: http://apache-kylin-incubating.74782.x6.nabble.com/Kylin-cube-build-fails-for-sample-cube-and-with-Hive-tables-tp2332.html Sent from the Apache Kylin (Incubating) mailing list archive at Nabble.com.
