[
https://issues.apache.org/jira/browse/HADOOP-3149?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12584401#action_12584401
]
Alejandro Abdelnur commented on HADOOP-3149:
--------------------------------------------
Owen, let me try filling the gaps in your suggestion.
# {{MultipleOutputFormats}} extends {{MultipleOutputFormat}} providing a static
method to define named outputs in the job conf including outputformat, key and
value classes.
# {{MultipleOutputFormats}} will also implement the logic to create the per
task file names based on the given named output.
# A {{MultipleOutputCollector}} will provide a {{collect(String namedOutput,
WritableComparable key, Writable value)}} method.
# {{MultipleOutputMapper}} and {{MultipleOutputReducer}} classes will have
{{map}} and {{reduce}} method receiving a {{MultipleOutputCollector}}.
# There will not be a {{MOJobConf}}, #1 takes care of seeding the named output
data into the vanilla {{JobConf}}.
Follow up comments/questions:
* I don't see the point of the KeyValue class as it will not be possible to
leverage the generic type checking at compile time (that is why #3 signature
differs from your suggestion).
* It seems to me that {{MultipleOutputFormats}} would have to have its own
subclasses for Text and SequenceFile to create the right {{RecordWriter}}
within the {{getRecordWriter}} method as I could have named outputs that are
{{TextOutputFormat}} or {{SequenceFileOutputFormat}}. And the
{{MultipleOutputFormats}} would be restricted to handle those 2 output format
types.
Thoughts?
> supporting multiple outputs for M/R jobs
> ----------------------------------------
>
> Key: HADOOP-3149
> URL: https://issues.apache.org/jira/browse/HADOOP-3149
> Project: Hadoop Core
> Issue Type: New Feature
> Components: mapred
> Environment: all
> Reporter: Alejandro Abdelnur
> Assignee: Alejandro Abdelnur
> Fix For: 0.17.0
>
> Attachments: patch3149.txt
>
>
> The outputcollector supports writing data to a single output, the 'part'
> files in the output path.
> We found quite common that our M/R jobs have to write data to different
> output. For example when classifying data as NEW, UPDATE, DELETE, NO-CHANGE
> to later do different processing on it.
> Handling the initialization of additional outputs from within the M/R code
> complicates the code and is counter intuitive with the notion of job
> configuration.
> It would be desirable to:
> # Configure the additional outputs in the jobconf, potentially specifying
> different outputformats, key and value classes for each one.
> # Write to the additional outputs in a similar way as data is written to the
> outputcollector.
> # Support the speculative execution semantics for the output files, only
> visible in the final output for promoted tasks.
> To support multiple outputs the following classes would be added to
> mapred/lib:
> * {{MOJobConf}} : extends {{JobConf}} adding methods to define named outputs
> (name, outputformat, key class, value class)
> * {{MOOutputCollector}} : extends {{OutputCollector}} adding a
> {{collect(String outputName, WritableComparable key, Writable value)}} method.
> * {{MOMapper}} and {{MOReducer}}: implement {{Mapper}} and {{Reducer}} adding
> a new {{configure}}, {{map}} and {{reduce}} signature that take the
> corresponding {{MO}} classes and performs the proper initialization.
> The data flow behavior would be: key/values written to the default (unnamed)
> output (using the original OutputCollector {{collect}} signature) take part
> of the shuffle/sort/reduce processing phases. key/values written to a named
> output from within a map don't.
> The named output files would be named using the task type and task ID to
> avoid collision among tasks (i.e. 'new-m-00002' and 'new-r-00001').
> Together with the setInputPathFilter feature introduced by HADOOP-2055 it
> would become very easy to chain jobs working on particular named outputs
> within a single directory.
> We are using heavily this pattern and it greatly simplified our M/R code as
> well as chaining different M/R.
> We wanted to contribute this back to Hadoop as we think is a generic feature
> many could benefit from.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
