[
https://issues.apache.org/jira/browse/HADOOP-1986?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12548401
]
Doug Cutting commented on HADOOP-1986:
--------------------------------------
Tom, this looks great to me. Bravo!
I think including at least one non-Writable serializer in the initial commit
would be best, to demonstrate the generality of the abstraction. Probably
JavaSerialization would be best, since it has no external dependencies.
Also, we don't need RecordSerialization do we, since records currently
implement Writable? However many would prefer it if records didn't implement
Writable. So a RecordSerialization that didn't rely on Writable but only on
the Record base class would be great to have. Then we could deperecate the
implementation of Writable by the record compiler, and make record io available
as a separate, standalone jar, as some have requested.
> Add support for a general serialization mechanism for Map Reduce
> ----------------------------------------------------------------
>
> Key: HADOOP-1986
> URL: https://issues.apache.org/jira/browse/HADOOP-1986
> Project: Hadoop
> Issue Type: New Feature
> Components: mapred
> Reporter: Tom White
> Assignee: Tom White
> Fix For: 0.16.0
>
> Attachments: hadoop-serializer-v2.tar.gz, SerializableWritable.java,
> serializer-v1.patch, serializer-v2.patch
>
>
> Currently Map Reduce programs have to use WritableComparable-Writable
> key-value pairs. While it's possible to write Writable wrappers for other
> serialization frameworks (such as Thrift), this is not very convenient: it
> would be nicer to be able to use arbitrary types directly, without explicit
> wrapping and unwrapping.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
