[
https://issues.apache.org/jira/browse/HADOOP-674?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12467919
]
Milind Bhandarkar commented on HADOOP-674:
------------------------------------------
One can easily use ByteArrayOutputStream to serialize and deserialize out of
current generated code. Addition of above API's will unnecessarily increase
complexity of generated code. Is there any strong reason to believe that the
above API will be a common case for usage ? In light of HADOOP-941, this API
will increase Hadoop Record I/O's dependence on Hadoop.
> Hadoop records should provide comvenient APIs for
> serialization/deserialization
> -------------------------------------------------------------------------------
>
> Key: HADOOP-674
> URL: https://issues.apache.org/jira/browse/HADOOP-674
> Project: Hadoop
> Issue Type: Improvement
> Components: record
> Reporter: Runping Qi
> Assigned To: Milind Bhandarkar
>
> I found the following APIs are very convenient to use and should be part of
> the generated class from Jute IDL :
> public BytesWritable serialize(String format);
> public void deserialize(BytesWitable data, String format);
> public static MyRecordJT deserialize(BytesWitable data, String format);
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
