[
https://issues.apache.org/jira/browse/PIG-3369?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13707461#comment-13707461
]
Johnny Zhang commented on PIG-3369:
-----------------------------------
turn out in hadoop23 profile, the test expects
{noformat}
assertEquals("(1,1,1001)", list.get(0).toString());
assertEquals("(3,3,1002)", list.get(1).toString());
assertEquals("(3,2,1002)", list.get(2).toString());
assertEquals("(3,1,1002)", list.get(3).toString());
assertEquals("(2,3,1002)", list.get(4).toString());
assertEquals("(2,2,1002)", list.get(5).toString());
assertEquals("(2,1,1002)", list.get(6).toString());
assertEquals("(1,3,1002)", list.get(7).toString());
assertEquals("(1,2,1002)", list.get(8).toString());
assertEquals("(1,1,1002)", list.get(9).toString());
assertEquals("(2,2,1003)", list.get(10).toString());
assertEquals("(2,1,1003)", list.get(11).toString());
assertEquals("(1,2,1003)", list.get(12).toString());
assertEquals("(1,1,1003)", list.get(13).toString());
assertEquals("(1,1,1004)", list.get(14).toString());
assertEquals("(3,3,1005)", list.get(15).toString());
assertEquals("(3,2,1005)", list.get(16).toString());
assertEquals("(2,3,1005)", list.get(17).toString());
assertEquals("(2,2,1005)", list.get(18).toString());
assertEquals("(3,3,1042)", list.get(19).toString());
{noformat}
instead of
{noformat}
assertEquals("(1,1,1001)", list.get(0).toString());
assertEquals("(1,1,1002)", list.get(1).toString());
assertEquals("(1,2,1002)", list.get(2).toString());
assertEquals("(1,3,1002)", list.get(3).toString());
assertEquals("(2,1,1002)", list.get(4).toString());
assertEquals("(2,2,1002)", list.get(5).toString());
assertEquals("(2,3,1002)", list.get(6).toString());
assertEquals("(3,1,1002)", list.get(7).toString());
assertEquals("(3,2,1002)", list.get(8).toString());
assertEquals("(3,3,1002)", list.get(9).toString());
assertEquals("(1,1,1003)", list.get(10).toString());
assertEquals("(1,2,1003)", list.get(11).toString());
assertEquals("(2,1,1003)", list.get(12).toString());
assertEquals("(2,2,1003)", list.get(13).toString());
assertEquals("(1,1,1004)", list.get(14).toString());
assertEquals("(2,2,1005)", list.get(15).toString());
assertEquals("(2,3,1005)", list.get(16).toString());
assertEquals("(3,2,1005)", list.get(17).toString());
assertEquals("(3,3,1005)", list.get(18).toString());
assertEquals("(3,3,1042)", list.get(19).toString());
{noformat}
the order is kind of reversed. The reason is because for same input (D1 and
D2), below command generate different results
{code}
J = JOIN D1 By shopId, D2 by shopId;
{code}
still looking for the root cause
> unit test TestImplicitSplitOnTuple.testImplicitSplitterOnTuple failed when
> using hadoopversion=23
> -------------------------------------------------------------------------------------------------
>
> Key: PIG-3369
> URL: https://issues.apache.org/jira/browse/PIG-3369
> Project: Pig
> Issue Type: Bug
> Affects Versions: 0.12
> Reporter: Johnny Zhang
>
> in trunk, this test pass with hadoopversion 20, but when using
> -Dhadoopversion=23, it fail. The error message is
> {noformat}
> expected:<([1,1],1002)> but was:<([3,3],1002)>
> {noformat}
--
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
