[ https://issues.apache.org/jira/browse/HADOOP-1986?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12537840 ]
Vivek Ratan commented on HADOOP-1986: ------------------------------------- Another problem I see is that _RecordSerializer_ only works for those classes that have constructors that take no arguments (otherwise _Class.newInstance()_ will fail, right?). This is also restrictive. As I mentioned in my previous comment, in platforms that require clients to pass in a class to deserialize, the classes can be quite complicated, and may not just be a wrapper around data. These classes need not have constructors that take in no arguments. Forcing a user to add such a constructor is unnecessarily restrictive. > 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, serializer-v1.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.