Kmeans mapred params in java

[Chris Harrington <ch...@heystaks.com>]

Hi all,

I've been trying to figure out how to set the following 2 hadoop params

-Dmapred.reduce.tasks=4 
-Dmapred.map.tasks=4

I understand that you set them via the command line 

bin/hadoop jar MyJar -Dmapred.reduce.tasks=4

but I've set MyJar to accept its own cmd line args so -Dmapred.reduce.tasks isn't getting passed to hadoop

Inside MyJar I call the KMeansDriver.run to do the clustering, so is there there a way to pass the above reduce/map task params programmatically to the Kmeans job.

I've been searching for a few hours now and I'm still coming up empty so any help would be appreciated.

Thanks