[
https://issues.apache.org/jira/browse/PIG-4042?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14046960#comment-14046960
]
Brian Johnson commented on PIG-4042:
------------------------------------
One difference I see between POMergeJoin and POMergeCogroup is that POMergeJoin
is schema aware and uses SchemaTupleBackend.newSchemaTupleFactory to create
it's tuple factory instead of the default tuple factory created in
POMergeCogroup.
> Casting isn't handled properly when using IndexableLoadFunc
> -----------------------------------------------------------
>
> Key: PIG-4042
> URL: https://issues.apache.org/jira/browse/PIG-4042
> Project: Pig
> Issue Type: Bug
> Affects Versions: 0.12.1
> Reporter: Brian Johnson
> Attachments: TEST-org.apache.pig.test.TestHBaseStorage.txt, patch
>
>
> I am working on a patch for HBaseStorage to implement IndexableLoadFunc. The
> results that are returned for the RHS of the join are always bytearrays
> regardless of the schema that is set, the caster, etc.. and the join only
> works if both relations have the row key set to be a bytearray. I have
> attached the patch I am working on with the two failing test cases. I am very
> eager to get this resolved, but after digging through the code I'm not sure
> where to start to fix it.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
