spark git commit: [SPARK-12026][MLLIB] ChiSqTest gets slower and slower over time when number of features is large

[jkbradley]

Repository: spark
Updated Branches:
  refs/heads/master cd81fc9e8 -> 021dafc6a


[SPARK-12026][MLLIB] ChiSqTest gets slower and slower over time when number of 
features is large

jira: https://issues.apache.org/jira/browse/SPARK-12026

The issue is valid as features.toArray.view.zipWithIndex.slice(startCol, 
endCol) becomes slower as startCol gets larger.

I tested on local and the change can improve the performance and the running 
time was stable.

Author: Yuhao Yang 

Closes #10146 from hhbyyh/chiSq.


Project: http://git-wip-us.apache.org/repos/asf/spark/repo
Commit: http://git-wip-us.apache.org/repos/asf/spark/commit/021dafc6
Tree: http://git-wip-us.apache.org/repos/asf/spark/tree/021dafc6
Diff: http://git-wip-us.apache.org/repos/asf/spark/diff/021dafc6

Branch: refs/heads/master
Commit: 021dafc6a05a31dc22c9f9110dedb47a1f913087
Parents: cd81fc9
Author: Yuhao Yang 
Authored: Wed Jan 13 17:43:27 2016 -0800
Committer: Joseph K. Bradley 
Committed: Wed Jan 13 17:43:27 2016 -0800

--
 .../scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala | 6 --
 1 file changed, 4 insertions(+), 2 deletions(-)
--


http://git-wip-us.apache.org/repos/asf/spark/blob/021dafc6/mllib/src/main/scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala
--
diff --git 
a/mllib/src/main/scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala 
b/mllib/src/main/scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala
index f22f2df..4a3fb064 100644
--- a/mllib/src/main/scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala
+++ b/mllib/src/main/scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala
@@ -109,7 +109,9 @@ private[stat] object ChiSqTest extends Logging {
   }
   i += 1
   distinctLabels += label
-  features.toArray.view.zipWithIndex.slice(startCol, endCol).map { 
case (feature, col) =>
+  val brzFeatures = features.toBreeze
+  (startCol until endCol).map { col =>
+val feature = brzFeatures(col)
 allDistinctFeatures(col) += feature
 (col, feature, label)
   }
@@ -122,7 +124,7 @@ private[stat] object ChiSqTest extends Logging {
   pairCounts.keys.filter(_._1 == 
startCol).map(_._3).toArray.distinct.zipWithIndex.toMap
   }
   val numLabels = labels.size
-  pairCounts.keys.groupBy(_._1).map { case (col, keys) =>
+  pairCounts.keys.groupBy(_._1).foreach { case (col, keys) =>
 val features = keys.map(_._2).toArray.distinct.zipWithIndex.toMap
 val numRows = features.size
 val contingency = new BDM(numRows, numLabels, new 
Array[Double](numRows * numLabels))


-
To unsubscribe, e-mail: commits-unsubscr...@spark.apache.org
For additional commands, e-mail: commits-h...@spark.apache.org

spark git commit: [SPARK-12026][MLLIB] ChiSqTest gets slower and slower over time when number of features is large

[jkbradley]

Repository: spark
Updated Branches:
  refs/heads/branch-1.6 26f13faa9 -> a490787da


[SPARK-12026][MLLIB] ChiSqTest gets slower and slower over time when number of 
features is large

jira: https://issues.apache.org/jira/browse/SPARK-12026

The issue is valid as features.toArray.view.zipWithIndex.slice(startCol, 
endCol) becomes slower as startCol gets larger.

I tested on local and the change can improve the performance and the running 
time was stable.

Author: Yuhao Yang 

Closes #10146 from hhbyyh/chiSq.

(cherry picked from commit 021dafc6a05a31dc22c9f9110dedb47a1f913087)
Signed-off-by: Joseph K. Bradley 


Project: http://git-wip-us.apache.org/repos/asf/spark/repo
Commit: http://git-wip-us.apache.org/repos/asf/spark/commit/a490787d
Tree: http://git-wip-us.apache.org/repos/asf/spark/tree/a490787d
Diff: http://git-wip-us.apache.org/repos/asf/spark/diff/a490787d

Branch: refs/heads/branch-1.6
Commit: a490787daa5ec11a5e30bc0df31f81edd54ccc6a
Parents: 26f13fa
Author: Yuhao Yang 
Authored: Wed Jan 13 17:43:27 2016 -0800
Committer: Joseph K. Bradley 
Committed: Wed Jan 13 17:43:38 2016 -0800

--
 .../scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala | 6 --
 1 file changed, 4 insertions(+), 2 deletions(-)
--


http://git-wip-us.apache.org/repos/asf/spark/blob/a490787d/mllib/src/main/scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala
--
diff --git 
a/mllib/src/main/scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala 
b/mllib/src/main/scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala
index 23c8d7c..1c583a4 100644
--- a/mllib/src/main/scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala
+++ b/mllib/src/main/scala/org/apache/spark/mllib/stat/test/ChiSqTest.scala
@@ -109,7 +109,9 @@ private[stat] object ChiSqTest extends Logging {
   }
   i += 1
   distinctLabels += label
-  features.toArray.view.zipWithIndex.slice(startCol, endCol).map { 
case (feature, col) =>
+  val brzFeatures = features.toBreeze
+  (startCol until endCol).map { col =>
+val feature = brzFeatures(col)
 allDistinctFeatures(col) += feature
 (col, feature, label)
   }
@@ -122,7 +124,7 @@ private[stat] object ChiSqTest extends Logging {
   pairCounts.keys.filter(_._1 == 
startCol).map(_._3).toArray.distinct.zipWithIndex.toMap
   }
   val numLabels = labels.size
-  pairCounts.keys.groupBy(_._1).map { case (col, keys) =>
+  pairCounts.keys.groupBy(_._1).foreach { case (col, keys) =>
 val features = keys.map(_._2).toArray.distinct.zipWithIndex.toMap
 val numRows = features.size
 val contingency = new BDM(numRows, numLabels, new 
Array[Double](numRows * numLabels))


-
To unsubscribe, e-mail: commits-unsubscr...@spark.apache.org
For additional commands, e-mail: commits-h...@spark.apache.org