[
https://issues.apache.org/jira/browse/HADOOP-1986?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12533622
]
Vivek Ratan commented on HADOOP-1986:
-------------------------------------
I forgot to add - using introspection to walk through a class has some more
problems. Things like versioning, and other kinds of optimizations will be
difficult with general introspection. As someone mentioned in the Thrift
discussion, it also doesn't handle field IDs in DDLs, and other features (such
as default values, optional fields) that serialization platforms will want to
support. For such cases, DDL-generated classes are better. By offering both
approaches, we can let users and developers decide on the tradeoff : ease of
use versus more complicated functionality.
> 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: SerializableWritable.java
>
>
> 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.
