[
https://issues.apache.org/jira/browse/AVRO-1353?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13728164#comment-13728164
]
Doug Cutting commented on AVRO-1353:
------------------------------------
In trunk (soon to be Avro 1.7.5) GenericData already serves the purpose of your
DatumMapping interface. In particular, for every DatumReader/DatumWriter
implementation there's a subclass of GenericData that implements
createDatumReader(), createDatumWriter() and compare().
Perhaps these methods would be better abstracted into an abstract base class
named org.apache.hadoop.data.DataModel. (Abstract base classes with many
methods are easier to evolve back-compatibly than interfaces, so I prefer them
when there's more than one or two methods.) Additional methods that might make
sense to add here are perhaps deepCopy() and hashCode(). Adding this might be
done as a separate cleanup issue and isn't required to extend the MapReduce
implementations as that can be done using GenericData in the meantime.
> Configurable Hadoop serialization in-memory representations
> -----------------------------------------------------------
>
> Key: AVRO-1353
> URL: https://issues.apache.org/jira/browse/AVRO-1353
> Project: Avro
> Issue Type: Improvement
> Components: java
> Affects Versions: 1.7.5
> Reporter: Marshall Bockrath-Vandegrift
> Attachments: avro-1353-2.patch, avro-1353.patch
>
>
> As discussed on the Avro Users mailing list [1], it would be useful to allow
> configuration of the DatumReader/Writer implementations used by Hadoop Avro
> serialization, especially for non-Java JVM languages.
> [1]
> http://mail-archives.apache.org/mod_mbox/avro-user/201305.mbox/%3C87ehdh14qn.fsf%40zeno.atl.damballa%3E
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
