From 40c884b3d74a1779be80974fc6fc926b0812813c Mon Sep 17 00:00:00 2001
From: Patrick Simianer <simianer@cl.uni-heidelberg.de>
Date: Tue, 27 Sep 2011 00:10:48 +0200
Subject: some updates for hadoop streaming

---
 dtrain/hstreaming/avgweights.rb | 27 ---------------------------
 1 file changed, 27 deletions(-)
 delete mode 100755 dtrain/hstreaming/avgweights.rb

(limited to 'dtrain/hstreaming/avgweights.rb')

diff --git a/dtrain/hstreaming/avgweights.rb b/dtrain/hstreaming/avgweights.rb
deleted file mode 100755
index d5cfaa4d..00000000
--- a/dtrain/hstreaming/avgweights.rb
+++ /dev/null
@@ -1,27 +0,0 @@
-#!/usr/bin/env ruby1.9.1
-
-
-STDIN.set_encoding 'utf-8'
-
-#shard_count_key = "__SHARD_COUNT__"
-
-w = {}
-c = {}
-w.default = 0
-c.default = 0
-while line = STDIN.gets
-  key, val = line.split /\t/
-  w[key] += val.to_f
-  c[key] += 1.0
-end
-
-#shard_count = w["__SHARD_COUNT__"]
-
-w.each_key { |k|
-  #if k == shard_count_key then next end
-  #if k == "__bias" then next end
-  puts "#{k}\t#{w[k]/c[k]}" #{w[k]/shard_count}"
-}
-
-#puts "#{shard_count_key}\t#{w[shard_count_key]}"
-
-- 
cgit v1.2.3