From 6c6e47e3cd4b5c717517609a19632e61c02b7d51 Mon Sep 17 00:00:00 2001
From: Ankur Dave <ankurdave@gmail.com>
Date: Fri, 9 Sep 2011 01:46:00 +0000
Subject: [PATCH] Use BufferedOutputStream in ShuffleMapTask
---
core/src/main/scala/spark/ShuffleMapTask.scala | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/core/src/main/scala/spark/ShuffleMapTask.scala b/core/src/main/scala/spark/ShuffleMapTask.scala
index b64401c2c7..eb6a5e2df3 100644
--- a/core/src/main/scala/spark/ShuffleMapTask.scala
+++ b/core/src/main/scala/spark/ShuffleMapTask.scala
@@ -1,5 +1,6 @@
package spark
+import java.io.BufferedOutputStream
import java.io.FileOutputStream
import java.io.ObjectOutputStream
import scala.collection.mutable.HashMap
@@ -26,7 +27,7 @@ extends DAGTask[String](stageId) with Logging {
val ser = SparkEnv.get.serializer.newInstance()
for (i <- 0 until numOutputSplits) {
val file = LocalFileShuffle.getOutputFile(dep.shuffleId, partition, i)
- val out = ser.outputStream(new FileOutputStream(file))
+ val out = ser.outputStream(new BufferedOutputStream(new FileOutputStream(file)))
buckets(i).foreach(pair => out.writeObject(pair))
// TODO: have some kind of EOF marker
out.close()
--
GitLab