[ https://issues.apache.org/jira/browse/HIVE-13282?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Matt McCline updated HIVE-13282: -------------------------------- Attachment: HIVE-13282.01.patch Work-In-Progress. Changed GroupByOperator.process to ignore the tag and just use 0. This is what was causing the ArrayIndexOutOfBoundsException. Is this the correct solution here? Also, experimented with deferring close on the SMBMapJoinOperator and CommonMergeJoinOperator (Tez) until all the parents have closed. This produces wrong results for the ORDER BY query (0 instead of 480) and the new GROUP BY query (0 instead of ???) in tez_smb_1.q > GroupBy and select operator encounter ArrayIndexOutOfBoundsException > -------------------------------------------------------------------- > > Key: HIVE-13282 > URL: https://issues.apache.org/jira/browse/HIVE-13282 > Project: Hive > Issue Type: Bug > Components: Query Processor > Affects Versions: 1.2.1, 2.0.0, 2.1.0 > Reporter: Vikram Dixit K > Assignee: Matt McCline > Attachments: HIVE-13282.01.patch, smb_fail_issue.patch, > smb_groupby.q, smb_groupby.q.out > > > The group by and select operators run into the ArrayIndexOutOfBoundsException > when they incorrectly initialize themselves with tag 0 but the incoming tag > id is different. > {code} > select count(*) from > (select rt1.id from > (select t1.key as id, t1.value as od from tab t1 group by key, value) rt1) vt1 > join > (select rt2.id from > (select t2.key as id, t2.value as od from tab_part t2 group by key, value) > rt2) vt2 > where vt1.id=vt2.id; > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)