[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13637524#comment-13637524 ] Namit Jain commented on HIVE-4106: -- [~ashutoshc], confirmed that the test still failed after HIVE-4371 SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Namit Jain Priority: Blocker Attachments: auto_sortmerge_join_12.q, hive.4106.1.patch, hive.4106.2.patch, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13637568#comment-13637568 ] Navis commented on HIVE-4106: - [~namit] Sorry, it was a different NPE. Fixed that. SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Namit Jain Priority: Blocker Attachments: auto_sortmerge_join_12.q, hive.4106.1.patch, hive.4106.2.patch, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13637591#comment-13637591 ] Ashutosh Chauhan commented on HIVE-4106: [~navis] Didn't get you. You mean this jira is a non-issue after your latest patch on HIVE-4371. SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Namit Jain Priority: Blocker Attachments: auto_sortmerge_join_12.q, hive.4106.1.patch, hive.4106.2.patch, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13637201#comment-13637201 ] Namit Jain commented on HIVE-4106: -- I think it would still fail - I havent't tried it, but am nearly sure. Basically, it assumes the left aliases and right aliases in a certain way, and that assumption has been changed by the join merge now. It only matters in this peculiar case - when the table being joined 'd' does not show up in the join condition -- should not be a common case. SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Namit Jain Priority: Blocker Attachments: auto_sortmerge_join_12.q, hive.4106.1.patch, hive.4106.2.patch, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13636443#comment-13636443 ] Ashutosh Chauhan commented on HIVE-4106: Isn't HIVE-4371 a proper fix for it? Does the test-case still fails after applying HIVE-4371? SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Namit Jain Priority: Blocker Attachments: auto_sortmerge_join_12.q, hive.4106.1.patch, hive.4106.2.patch, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13633835#comment-13633835 ] Namit Jain commented on HIVE-4106: -- [~ashutoshc], can you taje a look ? SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Namit Jain Priority: Blocker Attachments: auto_sortmerge_join_12.q, hive.4106.1.patch, hive.4106.2.patch, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13632672#comment-13632672 ] Namit Jain commented on HIVE-4106: -- [~vikram.dixit], can you confirm ? SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13632791#comment-13632791 ] Vikram Dixit K commented on HIVE-4106: -- I am going to try this and provide an update in a bit. Thanks Vikram. SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13633566#comment-13633566 ] Vikram Dixit K commented on HIVE-4106: -- Hi [~namit] I am still able to reproduce this issue. The testcase provided in this jira still produces this issue as well. Thanks Vikram. SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13633568#comment-13633568 ] Vikram Dixit K commented on HIVE-4106: -- {noformat} java.lang.ArrayIndexOutOfBoundsException: 1 at org.apache.hadoop.hive.ql.optimizer.AbstractSMBJoinProc.canConvertJoinToBucketMapJoin(AbstractSMBJoinProc.java:476) at org.apache.hadoop.hive.ql.optimizer.AbstractSMBJoinProc.canConvertJoinToSMBJoin(AbstractSMBJoinProc.java:431) at org.apache.hadoop.hive.ql.optimizer.SortedMergeJoinProc.process(SortedMergeJoinProc.java:46) at org.apache.hadoop.hive.ql.lib.DefaultRuleDispatcher.dispatch(DefaultRuleDispatcher.java:89) at org.apache.hadoop.hive.ql.lib.DefaultGraphWalker.dispatch(DefaultGraphWalker.java:87) at org.apache.hadoop.hive.ql.lib.DefaultGraphWalker.walk(DefaultGraphWalker.java:124) at org.apache.hadoop.hive.ql.lib.DefaultGraphWalker.startWalking(DefaultGraphWalker.java:101) at org.apache.hadoop.hive.ql.optimizer.SortedMergeBucketMapJoinOptimizer.transform(SortedMergeBucketMapJoinOptimizer.java:109) ... {noformat} SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13633748#comment-13633748 ] Namit Jain commented on HIVE-4106: -- https://issues.apache.org/jira/secure/attachment/12572229/auto_sortmerge_join_12.q Do you mean the above ? SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13633763#comment-13633763 ] Vikram Dixit K commented on HIVE-4106: -- Yes. When I run the auto_sortmerge_join_12.q on latest trunk, I see the above exception. I tried to see if my patch addresses the issue, but I am seeing a different issue with the latest trunk which I am looking into. If you have any suggestions, I would like to explore that route as well. Thanks Vikram. SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13633772#comment-13633772 ] Namit Jain commented on HIVE-4106: -- Let me take a look. SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13633793#comment-13633793 ] Namit Jain commented on HIVE-4106: -- The last query is wrong: explain select count(*) FROM bucket_small a JOIN bucket_medium b ON a.key = b.key JOIN bucket_big c ON c.key = b.key JOIN bucket_medium d ON c.key = b.key; Most probably, the intention was: explain select count(*) FROM bucket_small a JOIN bucket_medium b ON a.key = b.key JOIN bucket_big c ON c.key = b.key JOIN bucket_medium d ON c.key = d.key; Having said that, it should not die. SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13633799#comment-13633799 ] Namit Jain commented on HIVE-4106: -- https://reviews.facebook.net/D10293 SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, hive.4106.1.patch, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13629979#comment-13629979 ] Namit Jain commented on HIVE-4106: -- [~vikram.dixit], can you load the failing query ? SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13629994#comment-13629994 ] Vikram Dixit K commented on HIVE-4106: -- [~namit] {noformat} select i_item_id ,i_item_desc ,i_current_price from item i join inventory inv on (inv.inv_item_sk = i.i_item_sk) join date_dim d on (d.d_date_sk = inv.inv_date_sk) join store_sales ss on (ss.ss_item_sk = i.i_item_sk) where i_current_price between 62 and 62+30.0 and d_date between '2000-05-25' and '2000-07-27' and i_manufact_id in (129,270,821,423) and inv_quantity_on_hand between 100 and 500 group by i_item_id,i_item_desc,i_current_price order by i_item_id limit 100; {noformat} This is the TPC-DS benchmark query (scale 1 but this does not matter) where store_sales and inventory are sorted as follows: store_sales: sorted by ss_item_sk, partitioned on ss_sold_date, clustered by ss_item_sk inventory: partitioned by (inv_date string) clustered by (inv_item_sk) sorted by (inv_item_sk) item: non-partitioned, bucketed clustered by (i_item_sk) sorted by (i_item_sk) date_dim: non-partitioned, non-bucketed. SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13629996#comment-13629996 ] Namit Jain commented on HIVE-4106: -- This should not a mutli-way SMB join - this is due to HIVE-4167. Can you apply the patch for 4167 and check if this works ? SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Priority: Blocker Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13628849#comment-13628849 ] Namit Jain commented on HIVE-4106: -- I think this is a blocker for 0.11. [~vikram.dixit], do you have any testcase that fails ? SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13621055#comment-13621055 ] Ashutosh Chauhan commented on HIVE-4106: [~namit] / [~vikram.dixit] Any update on this one. Would be good to have this in 0.11 ? SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13593767#comment-13593767 ] Namit Jain commented on HIVE-4106: -- // check if the join operator encountered is a candidate for being converted // to a sort-merge join private NodeProcessor getCheckCandidateJoin() { return new NodeProcessor() { @Override public Object process(Node nd, StackNode stack, NodeProcessorCtx procCtx, Object... nodeOutputs) throws SemanticException { SortBucketJoinProcCtx smbJoinContext = (SortBucketJoinProcCtx)procCtx; JoinOperator joinOperator = (JoinOperator)nd; int size = stack.size(); if (!(stack.get(size-1) instanceof JoinOperator) || !(stack.get(size-2) instanceof ReduceSinkOperator)) { smbJoinContext.getRejectedJoinOps().add(joinOperator); return null; } // If any operator in the stack does not support a auto-conversion, this join should // not be converted. for (int pos = size -3; pos = 0; pos--) { Operator? extends OperatorDesc op = (Operator? extends OperatorDesc)stack.get(pos); if (!op.supportAutomaticSortMergeJoin()) { smbJoinContext.getRejectedJoinOps().add(joinOperator); return null; } } return null; } }; } It should be done above - file SortedMergeBucketMapJoinOptimizer.java. Can you try your testcase, and see if it being added to rejectJoinOps ? SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Attachments: HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13594311#comment-13594311 ] Vikram Dixit K commented on HIVE-4106: -- This test somehow feels awkward to me. It produces the issue but the test I have is slightly different. I am trying to come up with a better test for this. SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Attachments: auto_sortmerge_join_12.q, HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13592353#comment-13592353 ] Namit Jain commented on HIVE-4106: -- As we discussed in HIVE-3891, Namit Jain added a comment - 01/Mar/13 12:28 canConvertJoinToBucketMapJoin: The join should be rejected by: // This has already been inspected and rejected if (context.getRejectedJoinOps().contains(joinOp)) { return false; } Can you open a separate jira for this ? We should never run into this scenario. this needs to be debugged. Can you upload the faling testcase ? SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Attachments: HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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
[jira] [Commented] (HIVE-4106) SMB joins fail in multi-way joins
[ https://issues.apache.org/jira/browse/HIVE-4106?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13593021#comment-13593021 ] Vikram Dixit K commented on HIVE-4106: -- Hi Namit, The test case is not from our set of files/tables/TestCliDriver tests. I had to get lucky to catch this issue. The issue can be reproduced if we have a partitioned table with average size being in some range so that getBigTablePosition in the AvgPartitionSizeBasedBigTableSelectorForAutoSMJ returns a position not in the joinContext. This will cause the exception. In my setup, I got lucky to have such a partition combination. I am not sure how I can reproduce this issue with our current partitioned tables set in the TestCliDriver. I will try it with some combinations to see if it can be produced. If we say this type of join has to be rejected, we will not convert this join to an SMB Map-join even if it is possible right? That seems extreme. Also, I do not see any references to setRejectedJoinOps in the code either. Please let me know. Thanks Vikram. SMB joins fail in multi-way joins - Key: HIVE-4106 URL: https://issues.apache.org/jira/browse/HIVE-4106 Project: Hive Issue Type: Bug Components: Query Processor Affects Versions: 0.11.0 Reporter: Vikram Dixit K Assignee: Vikram Dixit K Attachments: HIVE-4106.patch I see array out of bounds exception in case of multi way smb joins. This is related to changes that went in as part of HIVE-3403. This issue has been discussed in HIVE-3891. -- 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