hadoop-common-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Stu Hood" <stuh...@webmail.us>
Subject RE: configuration for mappers?
Date Sat, 03 Nov 2007 22:15:53 GMT
Hi Chris,

A very comfortable way to pass configuration parameters to mappers is in your JobConf. By
using JobConf.set(String,String), you can set arbitrary parameters that will be distributed
to all mappers.

In order to fetch these parameters in your Mapper, you will need to override the MapReduceBase.configure(JobConf)
method. The default implementation of this method in MapReduceBase does nothing with the configuration,
but you can implement it to change the configuration of the mapper by .get(String)'ing the
parameters you stored and setting private variables, etc.

Good luck!

Thanks,
Stu


-----Original Message-----
From: Chris Dyer <redpony@umd.edu>
Sent: Saturday, November 3, 2007 5:42pm
To: hadoop-user@lucene.apache.org
Subject: configuration for mappers?

Hi hadoopers,

What is the recommended way to send configuration type information
from the job client down to the individual map and reduce processes
that are running?  I'd like to be able to specify a few options that
control the behavior of these processes, but I haven't been able to
find any documentation on how to do this.

Thanks for any help!

Chris



Mime
View raw message