You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hive.apache.org by "Eugene Koifman (JIRA)" <ji...@apache.org> on 2018/09/25 23:10:00 UTC

[jira] [Commented] (HIVE-16953) OrcRawRecordMerger.discoverOriginalKeyBounds issue if both split start and end are in the same stripe

    [ https://issues.apache.org/jira/browse/HIVE-16953?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16628034#comment-16628034 ] 

Eugene Koifman commented on HIVE-16953:
---------------------------------------

better long term fix is to make VectorizedOrcAcidRowBatchReader is used for both vectorized and non-vectorized reads and add a thin wrapper on top of it for non-vectorized reads that turns VRBs into rows

> OrcRawRecordMerger.discoverOriginalKeyBounds issue if both split start and end are in the same stripe
> -----------------------------------------------------------------------------------------------------
>
>                 Key: HIVE-16953
>                 URL: https://issues.apache.org/jira/browse/HIVE-16953
>             Project: Hive
>          Issue Type: Bug
>          Components: Transactions
>    Affects Versions: 1.0.0
>            Reporter: Eugene Koifman
>            Priority: Major
>
> if getOffset() and getMaxOffset() are inside
>     * the sames tripe - in this case we have minKey & isTail=false but rowLength is never set.
> don't know if we can ever have a split like that



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)