[
https://issues.apache.org/jira/browse/HIVE-18158?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16380345#comment-16380345
]
Gopal V commented on HIVE-18158:
--------------------------------
The patch looks fine - +1
There's some commented code in the hash function which probably needs to be
removed or uncommented.
{code}
+// result = 31 * result + (isDeleteEvent ? 1 : 0);
{code}
The delete event ordering is mostly irrelevant today and the lookup should work
fine even if the hash function does not use this boolean.
> Remove OrcRawRecordMerger.ReaderPairAcid.statementId
> ----------------------------------------------------
>
> Key: HIVE-18158
> URL: https://issues.apache.org/jira/browse/HIVE-18158
> Project: Hive
> Issue Type: Improvement
> Components: Transactions
> Affects Versions: 3.0.0
> Reporter: Eugene Koifman
> Assignee: Eugene Koifman
> Priority: Minor
> Attachments: HIVE-18158.01.patch, HIVE-18158.02.patch,
> HIVE-18158.03.patch
>
>
> * Need to get rid of this since we can always get this from the row
> itself in Acid 2.0.
> * For Acid 1.0, statementId == 0 in all deltas because both
> multi-statement txns and
> * Split Upate are only available in test mode so there is nothing can
> create a
> * deltas_x_x_M with M > 0.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
