[ 
https://issues.apache.org/jira/browse/AVRO-593?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13212819#comment-13212819
 ] 

Doug Cutting commented on AVRO-593:
-----------------------------------

I see a few choices:

1. org.apache.avro.{mapred,mapreduce,io,file,util}.  This is what the code on 
github does.  This would make the avro-mapred module contain things outside the 
org.apache.avro.mapred package, and splits Avro's io, file and util packages 
across multiple modules.

2. org.apache.avro.mapred.{mapreduce,io,file,util}.  This is what my patch 
does.  This is back-compatible and consistent with the module name, but places 
mapreduce under mapred, which is different than the Hadoop layout.

3. org.apache.avro.hadoop.{mapred,mapreduce,io,file,util}.  We'd rename the 
module to be avro-hadoop.  This would be incompatible but consistent with 
Hadoop.  For back-compatibility we might leave the mapred classes in their 
current package.

4. org.apache.avro.{mapred,mapreduce,mapred.io,mapred.file,mapred.util}.  This 
is back-compatible but includes a package that's not under the package of the 
module name.

Tom, are you advocating for (4)?  I'd be okay with that, I guess.

I'm also leaning towards moving AvroSequenceFile under org.apache.avro and 
adding just a shim base class into org.apache.hadoop.io that subclasses 
SequenceFile and makes public the bits we need.  That way if we get Hadoop to 
expose these bits the Avro API would not change.



                
> Avro mapreduce apis incompatible with hadoop 0.20.2
> ---------------------------------------------------
>
>                 Key: AVRO-593
>                 URL: https://issues.apache.org/jira/browse/AVRO-593
>             Project: Avro
>          Issue Type: Bug
>          Components: java
>    Affects Versions: 1.3.2, 1.3.3
>         Environment: Avro 1.3.3, Hadoop 0.20.2
>            Reporter: Steve Severance
>            Assignee: Garrett Wu
>         Attachments: AVRO-593.patch, AVRO-593.patch
>
>
> The avro api's for hadoop use the hadoop mapreduce api that has been 
> deprecated. A new avro mapreduce api should be implemented for hadoop 0.20 
> and higher.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: 
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to