[ https://issues.apache.org/jira/browse/HIVE-3652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13490470#comment-13490470 ]
Namit Jain commented on HIVE-3652: ---------------------------------- I was thinking more from the point of the current implementation. A backup task is per join operation currently. Thinking more about it, we can have a backup task (which can be a tree of tasks). It would be very difficult to fit the following in the current architecture. There are 10 dimension tables, 9 of them fit into memory and one of them dont. Perform a map-only join for the first 9, and then a regular backup join for the last one. I am not sure, if we want to optimize that. > Join optimization for star schema > --------------------------------- > > Key: HIVE-3652 > URL: https://issues.apache.org/jira/browse/HIVE-3652 > Project: Hive > Issue Type: Improvement > Components: Query Processor > Reporter: Amareshwari Sriramadasu > Assignee: Amareshwari Sriramadasu > > Currently, if we join one fact table with multiple dimension tables, it > results in multiple mapreduce jobs for each join with dimension table, > because join would be on different keys for each dimension. > Usually all the dimension tables will be small and can fit into memory and so > map-side join can used to join with fact table. > In this issue I want to look at optimizing such query to generate single > mapreduce job sothat mapper loads dimension tables into memory and joins with > fact table on different keys as well. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira