hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Guang Yang <gy...@millennialmedia.com>
Subject Re: Hadoop map-side join
Date Wed, 14 Nov 2012 17:56:29 GMT
Thanks for the answer, Chris. Is there a plan to make the fix available to
1.x in the near future?

- Guang

On 11/14/12 12:42 PM, "Chris Douglas" <cdouglas@apache.org> wrote:

>See https://issues.apache.org/jira/browse/MAPREDUCE-355 (not in 1.x
>series) -C
>On Tue, Nov 13, 2012 at 8:26 AM, Guang Yang <gyang@millennialmedia.com>
>> Hi,
>> I'm trying to use Hadoop map-side join in my application and wondering
>> anybody knows if there's a way to use it with the new Hadoop API
>> ("org.apache.hadoop.mapreduce.*") instead of the old Hadoop API
>> ("org.apache.hadoop.mapred.*"). The input format I'm trying to use for
>> join is "CompositeInputFormat", which is in the old API package and
>> like it expects everything (job configuration, input split, etc) to be
>> the old API too. This is a problem for me because I'm using the new API
>> create my map/reduce jobs so I can't just use "CompositeInputFormat" as
>> job's input format. I wonder if the only way to get the map-side join
>> is to use the old API to create map/reduce jobs. I appreciate any
>> regarding this issue.
>> Thanks,
>> Guang Yang

View raw message