From d13ec3324429919dcea549b00bae2e83ba51073c Mon Sep 17 00:00:00 2001
From: Peter Rudenko <petro.rudenko@gmail.com>
Date: Fri, 13 Feb 2015 14:41:44 +0200
Subject: [PATCH] [Ml] SPARK-5796 Don't transform data on a last estimator in
 Pipeline

If it's a last estimator in Pipeline there's no need to transform data, since there's no next stage that would consume this data.
---
 mllib/src/main/scala/org/apache/spark/ml/Pipeline.scala | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/mllib/src/main/scala/org/apache/spark/ml/Pipeline.scala b/mllib/src/main/scala/org/apache/spark/ml/Pipeline.scala
index bb291e6e1fd7d..5607ed21afe18 100644
--- a/mllib/src/main/scala/org/apache/spark/ml/Pipeline.scala
+++ b/mllib/src/main/scala/org/apache/spark/ml/Pipeline.scala
@@ -114,7 +114,9 @@ class Pipeline extends Estimator[PipelineModel] {
             throw new IllegalArgumentException(
               s"Do not support stage $stage of type ${stage.getClass}")
         }
-        curDataset = transformer.transform(curDataset, paramMap)
+        if (index < indexOfLastEstimator) {
+          curDataset = transformer.transform(curDataset, paramMap)
+        }
         transformers += transformer
       } else {
         transformers += stage.asInstanceOf[Transformer]