hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Barna Zsombor Klara (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (HIVE-16231) Parquet timestamp may be stored differently since HIVE-12767
Date Fri, 17 Mar 2017 14:54:41 GMT

     [ https://issues.apache.org/jira/browse/HIVE-16231?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Barna Zsombor Klara updated HIVE-16231:
---------------------------------------
    Attachment: HIVE-16231.01.patch

In the use case where we read/write timestamps in legacy mode a regression was introduced.
Instead of writing the timestamps adjusted with the local timezone it was written and read
with UTC adjustment, so with no adjustment.
This patch will address this regression and should be released together with HIVE-12767.

> Parquet timestamp may be stored differently since HIVE-12767
> ------------------------------------------------------------
>
>                 Key: HIVE-16231
>                 URL: https://issues.apache.org/jira/browse/HIVE-16231
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Barna Zsombor Klara
>            Assignee: Barna Zsombor Klara
>            Priority: Critical
>             Fix For: 2.2.0
>
>         Attachments: HIVE-16231.01.patch
>
>
> If the parquet table is missing its timezone property then the timestamp will be stored
with an adjustment instead of without it. This will cause a regression with other applications
like Impala or Spark.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message