[
https://issues.apache.org/jira/browse/HIVE-10220?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14481911#comment-14481911
]
Matt McCline commented on HIVE-10220:
-------------------------------------
+1
(I was concerned that MapJoinOperator was going to reload a hash table and
HybridHashTableContainer wouldn't use the right thread safe object. But it
appears it reloads into a MapJoinContainer because it is dealing with 1
reloaded hash table at a time.)
> Concurrent read access within HybridHashTableContainer
> -------------------------------------------------------
>
> Key: HIVE-10220
> URL: https://issues.apache.org/jira/browse/HIVE-10220
> Project: Hive
> Issue Type: Bug
> Affects Versions: 1.2.0
> Reporter: Gopal V
> Assignee: Gopal V
> Attachments: HIVE-10220.1.patch
>
>
> HybridHashTableContainer can end up being cached if it does not spill - that
> needs to follow HIVE-10128 thread safety patterns for the partitioned hash
> maps.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
