[
https://issues.apache.org/jira/browse/HIVE-20154?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vincent Tran updated HIVE-20154:
--------------------------------
Component/s: UDF
> Improve unix_timestamp(args) to handle automatic-DST switching timezones
> ------------------------------------------------------------------------
>
> Key: HIVE-20154
> URL: https://issues.apache.org/jira/browse/HIVE-20154
> Project: Hive
> Issue Type: Improvement
> Components: UDF
> Affects Versions: 1.1.0
> Reporter: Vincent Tran
> Priority: Major
>
> Currently unix_timestamp(args) UDF will only handle static timezone
> specifiers. It does not recognize SystemV specifiers such as EST5EDT or
> PST8PDT.
> Based on this experiment, when z is used to parse a TZ string like UTC4PDT
> (obviously not a valid SystemV specifier) - it will parse the time as UTC.
> When z<n>z is used to parse a TZ string like UTC4PDT, it will take parse the
> timestamp as the TZ of the final z position. This is demonstrated by my final
> query when the format string z4z1z is used to parse UTC4PDT1EDT.
> {noformat}
> 0: jdbc:hive2://localhost:10000/default>; select
> from_unixtime(unix_timestamp("2018-02-01 00:00:00 UTC4PDT", "yyyy-MM-dd
> HH:mm:ss z"), "yyyy-MM-dd HH:mm:ss zzzz");
> +--------------------------------------------+--+
> | _c0 |
> +--------------------------------------------+--+
> | 2018-01-31 16:00:00 Pacific Standard Time |
> +--------------------------------------------+--+
> 1 row selected (0.041 seconds)
> 0: jdbc:hive2://localhost:10000/default>; select
> from_unixtime(unix_timestamp("2018-02-01 00:00:00 UTC", "yyyy-MM-dd HH:mm:ss
> z"), "yyyy-MM-dd HH:mm:ss zzzz");
> +--------------------------------------------+--+
> | _c0 |
> +--------------------------------------------+--+
> | 2018-01-31 16:00:00 Pacific Standard Time |
> +--------------------------------------------+--+
> 1 row selected (0.041 seconds)
> 0: jdbc:hive2://localhost:10000/default>; select
> from_unixtime(unix_timestamp("2018-02-01 00:00:00 UTC4PDT", "yyyy-MM-dd
> HH:mm:ss z4z"), "yyyy-MM-dd HH:mm:ss zzzz");
> +--------------------------------------------+--+
> | _c0 |
> +--------------------------------------------+--+
> | 2018-01-31 23:00:00 Pacific Standard Time |
> +--------------------------------------------+--+
> 1 row selected (0.047 seconds)
> 0: jdbc:hive2://localhost:10000/default>; select
> from_unixtime(unix_timestamp("2018-02-01 00:00:00 UTC4PDT1EDT", "yyyy-MM-dd
> HH:mm:ss z4z1z"), "yyyy-MM-dd HH:mm:ss zzzz");
> +--------------------------------------------+--+
> | _c0 |
> +--------------------------------------------+--+
> | 2018-01-31 20:00:00 Pacific Standard Time |
> +--------------------------------------------+--+
> 1 row selected (0.055 seconds)
> 0: jdbc:hive2://localhost:10000/default>;
> {noformat}
> So all in all, I don't think the SystemV specifier EST5EDT or PST8PDT are
> valid to unix_timestamp(args) at all. And that those when parsed with the
> z<n>z format string, will be read as whatever valid timezone at the final
> position (effectively EDT and PDT respectively in when those valid SystemV TZ
> specifiers above are used).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
