[
https://issues.apache.org/jira/browse/HIVE-13680?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Kevin Liew updated HIVE-13680:
------------------------------
Attachment: HIVE-13680.patch
First patch submitted.
HiveConf settings submitted by the client are required to be prefixed by
"set:hiveconf" so in ThriftCLIService we have to preserve this prefix to ensure
that the SessionState is generated correctly.
To make the code cleaner, we could expose functions (which are currently out of
scope) to parse the prefix, or we could have the client send the list of
compressors and list of configs in new fields in the Thrift message (after
compressor negotiation, the CompDe would be stored in a new field in
SessionState instead of in SessionState.sessConf). I prefer the second option.
> HiveServer2: Provide a way to compress ResultSets
> -------------------------------------------------
>
> Key: HIVE-13680
> URL: https://issues.apache.org/jira/browse/HIVE-13680
> Project: Hive
> Issue Type: Sub-task
> Components: HiveServer2, JDBC
> Reporter: Vaibhav Gumashta
> Assignee: Kevin Liew
> Attachments: HIVE-13680.patch, proposal.pdf
>
>
> With HIVE-12049 in, we can provide an option to compress ResultSets before
> writing to disk. The user can specify a compression library via a config
> param which can be used in the tasks.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
