[
https://issues.apache.org/jira/browse/MAPREDUCE-1126?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12801022#action_12801022
]
Aaron Kimball commented on MAPREDUCE-1126:
------------------------------------------
Do you think that it's okay to add methods like {{Job.setMapOutputKeySchema()}}
then? In the limit, if another serialization framework makes its way into
common use (e.g., Hessian, Protobufs) in Hadoop, we would then need to add a
mechanism to set their serialization system-specific metadata to {{Job}} as
well. We factored out InputFormat/OutputFormat-specific getters and setters
(c.f. {{FileInputFormat.addInputPath()}}) a while back, and deprecated
{{JobConf.addInputPath()}}; this seems like a logical next step.
Furthermore, what specific framework dependencies are you referring to? The
jobdata package provides getters and setters that allow users to configure
serialization system-specific metadata keys and values, but they are put into
well-defined "system wide" metadata locations (e.g.
{{JobContext.MAP_OUTPUT_KEY_METADATA}}) in the Configuration itself. The
SerializerBase/DeserializerBase classes are instantiated in JobConf without
touching the {{jobdata}} package at all (they rely only on the system-wide
Configuration names).
The only dependency on {{jobdata}} classes in Job/JobContext/JobConf is to
push-down the now-deprecated getter/setter methods that the user would call in
legacy code, but the framework no longer makes any calls to
{{JobConf.getMapOutputKeyClass()}}. It instead calls
{{JobConf.getMapOutputKeySerializer()}} and
{{JobConf.getMapOutputKeyDeserializer()}} directly.
> shuffle should use serialization to get comparator
> --------------------------------------------------
>
> Key: MAPREDUCE-1126
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-1126
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Components: task
> Reporter: Doug Cutting
> Assignee: Aaron Kimball
> Fix For: 0.22.0
>
> Attachments: MAPREDUCE-1126.2.patch, MAPREDUCE-1126.3.patch,
> MAPREDUCE-1126.4.patch, MAPREDUCE-1126.5.patch, MAPREDUCE-1126.6.patch,
> MAPREDUCE-1126.patch
>
>
> Currently the key comparator is defined as a Java class. Instead we should
> use the Serialization API to create key comparators. This would permit,
> e.g., Avro-based comparators to be used, permitting efficient sorting of
> complex data types without having to write a RawComparator in Java.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
