map reduce examples on HBaase
Java
Latest commit 97db6b3 Apr 11, 2010 @sujee readme update
Permalink
Failed to load latest commit information.
.settings eclipse project files Apr 9, 2010
src/hbase_mapred1 refactor, tweaks Apr 10, 2010
.classpath eclipse project files Apr 9, 2010
.gitignore refactor, tweaks Apr 10, 2010
.project eclipse project files Apr 9, 2010
README.txt readme update Apr 10, 2010

README.txt

for detailed tutorial : 
http://sujee.net/tech/articles/hbase-map-reduce-freq-counter/


To compile the project:

1) open in Eclisle
2) set HBASE_HOME  class variable to hbase install dir


setup hbase tables:
---
open hbase shell
	$ hbase shell
	  	create 'access_logs', 'details'
	  	create 'summary_user', {NAME=>'details', VERSIONS=>1}
	  	
'access_logs' is the 'raw' logs.  The key is userID+counter  (int + int)
'summary_user' is to compute summary.  key is 'userID' (int)

Running map reduce
--
1) run 'FreqCounter1' directly from Eclipse, as a Java application

2) run on cluster / command line
  a) make a jar
      jar cf freqCounter.jar -C classes .

  b) hadoop jar freqCounter.jar hbase_mapred1.FreqCounter1
      check progress at task tracker : http://localhost:50070