[ https://issues.apache.org/jira/browse/HIVE-262?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12669286#action_12669286 ]
Namit Jain commented on HIVE-262: --------------------------------- Ashish suggested the following approach: Based on join conditions, create a set of all tables being joined (no outer join), and if one of them is null for a given value, all of them become null. For example, A join B on A.c1=B.c1 join C on A.c1=C.c1 right outer join D on A.c1=D.c1 A,B,C belong to the same group (since A joins with B and A joins with C). So, for a given key (c1), if there is no row corresponding to either of A, B, or C - assume that there is no row for all of them for that key. That works for the example above, and the approach is different from the patch > outer join gets some duplicate rows in some scenarios > ----------------------------------------------------- > > Key: HIVE-262 > URL: https://issues.apache.org/jira/browse/HIVE-262 > Project: Hadoop Hive > Issue Type: Bug > Components: Query Processor > Affects Versions: 0.2.0 > Reporter: Namit Jain > Assignee: Namit Jain > Priority: Blocker > Fix For: 0.2.0 > > Attachments: patch.262.1.txt, patch262.2.txt > > > SELECT * FROM src src1 JOIN src src2 ON (src1.key = src2.key AND src1.key < > 10) RIGHT OUTER JOIN src src3 ON (src1.key = src3.key AND src3.key < 20); > returns duplicate rows for outer join -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.