From d947cb97789acd04b7f02410490fa3f91c2a7bca Mon Sep 17 00:00:00 2001
From: Matei Zaharia <matei@eecs.berkeley.edu>
Date: Wed, 3 Nov 2010 22:52:41 -0700
Subject: [PATCH] Fixed a bug with negative hashcodes
---
src/scala/spark/DfsShuffle.scala | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/src/scala/spark/DfsShuffle.scala b/src/scala/spark/DfsShuffle.scala
index a100ddf05b..e751e1bd75 100644
--- a/src/scala/spark/DfsShuffle.scala
+++ b/src/scala/spark/DfsShuffle.scala
@@ -84,7 +84,10 @@ extends Logging
new ObjectOutputStream(fs.create(path, true))
}).toArray
for ((k, c) <- combiners) {
- val bucket = k.hashCode % numOutputSplits
+ var bucket = k.hashCode % numOutputSplits
+ if (bucket < 0) {
+ bucket += numOutputSplits
+ }
outputStreams(bucket).writeObject((k, c))
}
outputStreams.foreach(_.close())
--
GitLab