(Bcc: Tez, Cross-post to hive) > I added ³set hive.execution.engine=mr;² at top of the script, seems the >result is correct
Pretty sure it's due to the same table aliases for both dummy tables (they're both called _dummy_table) auto join conversion. hive> set hive.auto.convert.join=false; Should go back to using slower tagged joins even in Tez, which will add a table-tag i.e first table will be (<table-key>, 0) amd 2nd table will be (<table-key>, 1). I suspect the difference between the MR and Tez runs are lookup between the table-name + expr (both equal for _dummy_table.11). > per Jeff Zhang's thinking if you were to set the exec engine to 'mr' >would it still fail? if so, then its not Tez . :) Hive has a a whole set of join algorithms which can only work on Tez, so it's not always that easy. Considering this is on hive-2.0.0, I recommend filing a JIRA on 2.0.0 and marking it with 2.0.1 as a target version. Cheers, Gopal
