From b6a8fdc981daa3b01826a3be4c2355ee5cd61f0a Mon Sep 17 00:00:00 2001
From: Patrick Simianer <p@simianer.de>
Date: Mon, 28 Nov 2011 11:34:41 +0100
Subject: hstreaming stuff

---
 dtrain/hstreaming/red-avg.rb | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

(limited to 'dtrain/hstreaming/red-avg.rb')

diff --git a/dtrain/hstreaming/red-avg.rb b/dtrain/hstreaming/red-avg.rb
index 771f4c0e..c0b69eb4 100755
--- a/dtrain/hstreaming/red-avg.rb
+++ b/dtrain/hstreaming/red-avg.rb
@@ -1,6 +1,5 @@
 #!/usr/bin/env ruby1.9.1
 
-
 shard_count_key = "__SHARD_COUNT__"
 
 STDIN.set_encoding 'utf-8'
@@ -8,7 +7,7 @@ STDOUT.set_encoding 'utf-8'
 
 w = {}
 c = {}
-w.default = 0
+w.default = 0.
 c.default = 0
 while line = STDIN.gets
   key, val = line.split /\t/
@@ -19,7 +18,10 @@ end
 puts "# dtrain reducer: average"
 shard_count = w["__SHARD_COUNT__"]
 w.each_key { |k|
-  if k == shard_count_key then next end
-  puts "#{k}\t#{w[k]/shard_count}"
+  if k == shard_count_key
+    puts "# shard count: #{shard_count.to_i}"
+  else
+    puts "#{k}\t#{w[k]/shard_count}\t# #{c[k]}"
+  end
 }
 
-- 
cgit v1.2.3