[
https://issues.apache.org/jira/browse/HADOOP-2329?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12548012
]
Jim Kellerman commented on HADOOP-2329:
---------------------------------------
-1
Adding value types to the HBase server side is not a good idea. Since members
of a column family can be created on an ad-hoc basis, there would be a lot of
bookkeeping to do to determine if the family member should be of a particular
type. And if there were no data about a particular family member, what type
should it be?
It would be unacceptable to force all members of a column family to be the same
type.
Additionally there have been requests to loosen the restrictions on the row key
being a Text and instead, accepting any WritableComparable as the row key.
HADOOP-2197 would permit applications to tag columns with arbitrary key/value
pairs. Thus an application could store family member/type information using
this mechanism.
Bigtable is typeless and I think HBase should be as well.
> [Hbase Shell] Addition of Built-In Value Data Types for efficient accessing
> and stroing data
> --------------------------------------------------------------------------------------------
>
> Key: HADOOP-2329
> URL: https://issues.apache.org/jira/browse/HADOOP-2329
> Project: Hadoop
> Issue Type: New Feature
> Components: contrib/hbase
> Affects Versions: 0.16.0
> Reporter: Edward Yoon
> Assignee: Edward Yoon
> Fix For: 0.16.0
>
>
> A built-in data type is a fundamental data type that the hbase shell defines.
> (character strings, scalars, ranges, arrays, ... , etc)
> If you need a specialized data type that is not currently provided as a
> built-in type,
> you are encouraged to write your own user-defined data type using UDC(not yet
> implemented).
> (or contribute it for distribution in a future release of hbase shell)
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
