[
https://issues.apache.org/jira/browse/HIVE-1228?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12864582#action_12864582
]
John Sichi commented on HIVE-1228:
----------------------------------
Hmmm, scratch the :timestamp thing for now, since timestamps are cell-level
rather than row-level. I guess we could do :timestamp:column-name to map them
individually, but that deserves its own entry in JIRA.
> allow HBase key column to be anywhere in Hive table, and expose timestamp
> column for read and write
> ---------------------------------------------------------------------------------------------------
>
> Key: HIVE-1228
> URL: https://issues.apache.org/jira/browse/HIVE-1228
> Project: Hadoop Hive
> Issue Type: Improvement
> Components: HBase Handler
> Affects Versions: 0.6.0
> Reporter: John Sichi
> Assignee: John Sichi
> Fix For: 0.6.0
>
>
> Current column mapping support requires key column to be first in Hive table
> (and implicit rather than explicit in column mapping). We should allow it to
> be specified explicitly at any position within column mapping, e.g. with
> syntax :key to distinguish it from normal columns/families.
> Together with this, allow :timestamp column to be mapped. For writes, this
> would allow timestamp to be explicitly provided by query (although we may
> want to prevent this in some cases; see for example HIVE-1223 and HIVE-1225).
> For reads, this would allow timestamp to be fetched, and filters on
> timestamp could be used to find particular row versions. (If no filter is
> supplied, should we return all row versions or only the default latest
> version? The semantics need to be well-defined here.)
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
