[
https://issues.apache.org/jira/browse/MAPREDUCE-326?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12833276#action_12833276
]
Tom White commented on MAPREDUCE-326:
-------------------------------------
bq. That is completely untrue. It is mark unstable as a warning to users. I
would reject any incompatible change to the API. We should mark the base
classes (Mapper, Reducer, InputFormat, OutputFormat, Partitioner, RecordReader,
RecordWriter, Partitioner, *Context) as stable. Pig is already using the new
APIs and so are a lot of other business critical applications.
Sorry, there must some misunderstanding here. I meant the *low-level* API is
marked "unstable". The existing MapReduce APIs can and should be marked as
"stable". To be clear: this proposal does not change these APIs (Goal 2 in the
document).
bq. But it is the primary interface that our users use. If you are proposing
that we make this the "real" interface, then I've already expressed my -1.
Users don't want to think in bytes.
Again, I seem to have expressed myself unclearly, and I apologize if that's the
case. I am not proposing that we make the low-level API a user-level interface.
I agree that users don't want to think in bytes, but this proposal doesn't
expect them to. Users would use the same object-based MapReduce API they use
today, oblivious to the existence of a raw, low-level API.
> The lowest level map-reduce APIs should be byte oriented
> --------------------------------------------------------
>
> Key: MAPREDUCE-326
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-326
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Reporter: eric baldeschwieler
> Attachments: MAPREDUCE-326-api.patch, MAPREDUCE-326.pdf
>
>
> As discussed here:
> https://issues.apache.org/jira/browse/HADOOP-1986#action_12551237
> The templates, serializers and other complexities that allow map-reduce to
> use arbitrary types complicate the design and lead to lots of object creates
> and other overhead that a byte oriented design would not suffer. I believe
> the lowest level implementation of hadoop map-reduce should have byte string
> oriented APIs (for keys and values). This API would be more performant,
> simpler and more easily cross language.
> The existing API could be maintained as a thin layer on top of the leaner API.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
