U can try it using Distributed cache
In Driver
FileStatus[] list = fs.globStatus(extrafile);
for (FileStatus status : list) {
DistributedCache.addCacheFile(status.getPath().toUri(), conf);
}
In Map
URI[] cacheFiles = DistributedCache.getCacheFiles(conf);
Path getPath = new Path(cacheFiles[0].ge
Hi
You can directly use this right?
FileInputFormat.setInputPaths(job,new Path(args[0]));
FileOutputFormat.setOutputPath(job, new Path(args[1]));
Or you need extra input file to feed into mapper?
On Fri, Jun 20, 2014 at 11:57 AM, Shrivastava, Himnshu (GE Global Research,
Non-GE) wrote:
> Ho
you can get the input from some source(e.g. files) in the mapper setup()
method and emit it to the context.write() so that it can reach to the
reducer.
Raj K Singh
http://in.linkedin.com/in/rajkrrsingh
http://www.rajkrrsingh.blogspot.com
Mobile Tel: +91 (0
That’s exactly what MapReduce does. The input is processed by the mapper
function, and its output will be automatically sent into the reducer function.
Between mappers and reducers we have the automatic shuffle phase which sends
records with identical keys into one reducer call.
If you want to
This is what we do. I'm sorry I didn't quite this.
We read in the data through Mapper, do some operations on it, and pass the
Mapper output on to the Reducer. If you intend to just pass the data as it
is then just context.write it without doing anything else.
*Warm regards,*
*Mohammad Tariq*
*clo
