[
https://issues.apache.org/jira/browse/HIVE-12195?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15117751#comment-15117751
]
Ryan Blue commented on HIVE-12195:
----------------------------------
Thanks for fixing that, [~sinchii]. I just noticed in this version that the TZ
set is a transient instance field. I don't think there is a need for each
instance to have its own time zone set since those are constant for the life of
the JVM. Maybe this should be a static final field? I'm also wondering why it
is transient: if this instance is going to be serialized, then the TZ set
should be also (if it isn't static). One more minor thing: I generally use the
Set interface for variable types rather than the concrete class HashSet.
> Unknown zones should cause an error instead of silently failing
> ---------------------------------------------------------------
>
> Key: HIVE-12195
> URL: https://issues.apache.org/jira/browse/HIVE-12195
> Project: Hive
> Issue Type: Sub-task
> Components: Hive
> Reporter: Ryan Blue
> Assignee: Shinichi Yamashita
> Attachments: HIVE-12195.1.patch, HIVE-12195.2.patch
>
>
> Using an unknown time zone with the {{from_utc_timestamp}} or
> {{to_utc_timetamp}} methods returns the time un-adjusted instead of throwing
> an error:
> {code}
> hive> select from_utc_timestamp('2015-04-11 12:24:34.535', 'panda');
> OK
> 2015-04-11 12:24:34.535
> {code}
> This should be an error because users may attempt to adjust to valid but
> unknown zones, like PDT or MDT. This would produce incorrect results with no
> warning or error.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
