[SPARK-21278][PYSPARK] Upgrade to Py4J 0.10.6

dongjoon-hyun · holdenk · commit c8d0aba198c0 · 2017-07-05T16:33:23.000-07:00
## What changes were proposed in this pull request? This PR aims to bump Py4J in order to fix the following float/double bug. Py4J 0.10.5 fixes this (py4j/py4j#272) and the latest Py4J is 0.10.6. **BEFORE** ``` >>> df = spark.range(1) >>> df.select(df['id'] + 17.133574204226083).show() +--------------------+ |(id + 17.1335742042)| +--------------------+ | 17.1335742042| +--------------------+ ``` **AFTER** ``` >>> df = spark.range(1) >>> df.select(df['id'] + 17.133574204226083).show() +-------------------------+ |(id + 17.133574204226083)| +-------------------------+ | 17.133574204226083| +-------------------------+ ``` ## How was this patch tested? Manual. Author: Dongjoon Hyun <dongjoon@apache.org> Closes apache#18546 from dongjoon-hyun/SPARK-21278.
diff --git a/LICENSE b/LICENSE
@@ -263,7 +263,7 @@ The text of each license is also included at licenses/LICENSE-[project].txt.
      (New BSD license) Protocol Buffer Java API (org.spark-project.protobuf:protobuf-java:2.4.1-shaded - http://code.google.com/p/protobuf)
      (The BSD License) Fortran to Java ARPACK (net.sourceforge.f2j:arpack_combined_all:0.1 - http://f2j.sourceforge.net)
      (The BSD License) xmlenc Library (xmlenc:xmlenc:0.52 - http://xmlenc.sourceforge.net)
-     (The New BSD License) Py4J (net.sf.py4j:py4j:0.10.4 - http://py4j.sourceforge.net/)
+     (The New BSD License) Py4J (net.sf.py4j:py4j:0.10.6 - http://py4j.sourceforge.net/)
      (Two-clause BSD-style license) JUnit-Interface (com.novocode:junit-interface:0.10 - http://github.com/szeiger/junit-interface/)
      (BSD licence) sbt and sbt-launch-lib.bash
      (BSD 3 Clause) d3.min.js (https://github.com/mbostock/d3/blob/master/LICENSE)
diff --git a/bin/pyspark b/bin/pyspark
@@ -57,7 +57,7 @@ export PYSPARK_PYTHON
 
 # Add the PySpark classes to the Python path:
 export PYTHONPATH="${SPARK_HOME}/python/:$PYTHONPATH"
-export PYTHONPATH="${SPARK_HOME}/python/lib/py4j-0.10.4-src.zip:$PYTHONPATH"
+export PYTHONPATH="${SPARK_HOME}/python/lib/py4j-0.10.6-src.zip:$PYTHONPATH"
 
 # Load the PySpark shell.py script when ./pyspark is used interactively:
 export OLD_PYTHONSTARTUP="$PYTHONSTARTUP"
diff --git a/bin/pyspark2.cmd b/bin/pyspark2.cmd
@@ -30,7 +30,7 @@ if "x%PYSPARK_DRIVER_PYTHON%"=="x" (
 )
 
 set PYTHONPATH=%SPARK_HOME%\python;%PYTHONPATH%
-set PYTHONPATH=%SPARK_HOME%\python\lib\py4j-0.10.4-src.zip;%PYTHONPATH%
+set PYTHONPATH=%SPARK_HOME%\python\lib\py4j-0.10.6-src.zip;%PYTHONPATH%
 
 set OLD_PYTHONSTARTUP=%PYTHONSTARTUP%
 set PYTHONSTARTUP=%SPARK_HOME%\python\pyspark\shell.py
diff --git a/core/pom.xml b/core/pom.xml
@@ -335,7 +335,7 @@
     <dependency>
       <groupId>net.sf.py4j</groupId>
       <artifactId>py4j</artifactId>
-      <version>0.10.4</version>
+      <version>0.10.6</version>
     </dependency>
     <dependency>
       <groupId>org.apache.spark</groupId>
diff --git a/core/src/main/scala/org/apache/spark/api/python/PythonUtils.scala b/core/src/main/scala/org/apache/spark/api/python/PythonUtils.scala
@@ -32,7 +32,7 @@ private[spark] object PythonUtils {
     val pythonPath = new ArrayBuffer[String]
     for (sparkHome <- sys.env.get("SPARK_HOME")) {
       pythonPath += Seq(sparkHome, "python", "lib", "pyspark.zip").mkString(File.separator)
-      pythonPath += Seq(sparkHome, "python", "lib", "py4j-0.10.4-src.zip").mkString(File.separator)
+      pythonPath += Seq(sparkHome, "python", "lib", "py4j-0.10.6-src.zip").mkString(File.separator)
     }
     pythonPath ++= SparkContext.jarOfObject(this)
     pythonPath.mkString(File.pathSeparator)
diff --git a/dev/deps/spark-deps-hadoop-2.6 b/dev/deps/spark-deps-hadoop-2.6
@@ -156,7 +156,7 @@ parquet-jackson-1.8.2.jar
 pmml-model-1.2.15.jar
 pmml-schema-1.2.15.jar
 protobuf-java-2.5.0.jar
-py4j-0.10.4.jar
+py4j-0.10.6.jar
 pyrolite-4.13.jar
 scala-compiler-2.11.8.jar
 scala-library-2.11.8.jar
diff --git a/dev/deps/spark-deps-hadoop-2.7 b/dev/deps/spark-deps-hadoop-2.7
@@ -157,7 +157,7 @@ parquet-jackson-1.8.2.jar
 pmml-model-1.2.15.jar
 pmml-schema-1.2.15.jar
 protobuf-java-2.5.0.jar
-py4j-0.10.4.jar
+py4j-0.10.6.jar
 pyrolite-4.13.jar
 scala-compiler-2.11.8.jar
 scala-library-2.11.8.jar
diff --git a/python/README.md b/python/README.md
@@ -29,4 +29,4 @@ The Python packaging for Spark is not intended to replace all of the other use c
 
 ## Python Requirements
 
-At its core PySpark depends on Py4J (currently version 0.10.4), but additional sub-packages have their own requirements (including numpy and pandas).
+At its core PySpark depends on Py4J (currently version 0.10.6), but additional sub-packages have their own requirements (including numpy and pandas).
diff --git a/python/docs/Makefile b/python/docs/Makefile
@@ -7,7 +7,7 @@ SPHINXBUILD   ?= sphinx-build
 PAPER         ?=
 BUILDDIR      ?= _build
 
-export PYTHONPATH=$(realpath ..):$(realpath ../lib/py4j-0.10.4-src.zip)
+export PYTHONPATH=$(realpath ..):$(realpath ../lib/py4j-0.10.6-src.zip)
 
 # User-friendly check for sphinx-build
 ifeq ($(shell which $(SPHINXBUILD) >/dev/null 2>&1; echo $$?), 1)
diff --git a/python/lib/py4j-0.10.4-src.zip b/python/lib/py4j-0.10.4-src.zip
diff --git a/python/lib/py4j-0.10.6-src.zip b/python/lib/py4j-0.10.6-src.zip
diff --git a/python/setup.py b/python/setup.py
@@ -194,7 +194,7 @@ def _supports_symlinks():
             'pyspark.examples.src.main.python': ['*.py', '*/*.py']},
         scripts=scripts,
         license='http://www.apache.org/licenses/LICENSE-2.0',
-        install_requires=['py4j==0.10.4'],
+        install_requires=['py4j==0.10.6'],
         setup_requires=['pypandoc'],
         extras_require={
             'ml': ['numpy>=1.7'],
diff --git a/resource-managers/yarn/src/main/scala/org/apache/spark/deploy/yarn/Client.scala b/resource-managers/yarn/src/main/scala/org/apache/spark/deploy/yarn/Client.scala
@@ -1124,7 +1124,7 @@ private[spark] class Client(
         val pyArchivesFile = new File(pyLibPath, "pyspark.zip")
         require(pyArchivesFile.exists(),
           s"$pyArchivesFile not found; cannot run pyspark application in YARN mode.")
-        val py4jFile = new File(pyLibPath, "py4j-0.10.4-src.zip")
+        val py4jFile = new File(pyLibPath, "py4j-0.10.6-src.zip")
         require(py4jFile.exists(),
           s"$py4jFile not found; cannot run pyspark application in YARN mode.")
         Seq(pyArchivesFile.getAbsolutePath(), py4jFile.getAbsolutePath())
diff --git a/resource-managers/yarn/src/test/scala/org/apache/spark/deploy/yarn/YarnClusterSuite.scala b/resource-managers/yarn/src/test/scala/org/apache/spark/deploy/yarn/YarnClusterSuite.scala
@@ -249,7 +249,7 @@ class YarnClusterSuite extends BaseYarnClusterSuite {
     // needed locations.
     val sparkHome = sys.props("spark.test.home")
     val pythonPath = Seq(
-        s"$sparkHome/python/lib/py4j-0.10.4-src.zip",
+        s"$sparkHome/python/lib/py4j-0.10.6-src.zip",
         s"$sparkHome/python")
     val extraEnvVars = Map(
       "PYSPARK_ARCHIVES_PATH" -> pythonPath.map("local:" + _).mkString(File.pathSeparator),
diff --git a/sbin/spark-config.sh b/sbin/spark-config.sh
@@ -28,6 +28,6 @@ export SPARK_CONF_DIR="${SPARK_CONF_DIR:-"${SPARK_HOME}/conf"}"
 # Add the PySpark classes to the PYTHONPATH:
 if [ -z "${PYSPARK_PYTHONPATH_SET}" ]; then
   export PYTHONPATH="${SPARK_HOME}/python:${PYTHONPATH}"
-  export PYTHONPATH="${SPARK_HOME}/python/lib/py4j-0.10.4-src.zip:${PYTHONPATH}"
+  export PYTHONPATH="${SPARK_HOME}/python/lib/py4j-0.10.6-src.zip:${PYTHONPATH}"
   export PYSPARK_PYTHONPATH_SET=1
 fi

Original file line number	Diff line number	Diff line change
`@@ -30,7 +30,7 @@ if "x%PYSPARK_DRIVER_PYTHON%"=="x" (`
`30`	`30`	`)`
`31`	`31`
`32`	`32`	`set PYTHONPATH=%SPARK_HOME%\python;%PYTHONPATH%`
`33`		`-set PYTHONPATH=%SPARK_HOME%\python\lib\py4j-0.10.4-src.zip;%PYTHONPATH%`
	`33`	`+set PYTHONPATH=%SPARK_HOME%\python\lib\py4j-0.10.6-src.zip;%PYTHONPATH%`
`34`	`34`
`35`	`35`	`set OLD_PYTHONSTARTUP=%PYTHONSTARTUP%`
`36`	`36`	`set PYTHONSTARTUP=%SPARK_HOME%\python\pyspark\shell.py`
Original file line number	Diff line number	Diff line change
`@@ -32,7 +32,7 @@ private[spark] object PythonUtils {`
`32`	`32`	`val pythonPath = new ArrayBuffer[String]`
`33`	`33`	`for (sparkHome <- sys.env.get("SPARK_HOME")) {`
`34`	`34`	`pythonPath += Seq(sparkHome, "python", "lib", "pyspark.zip").mkString(File.separator)`
`35`		`- pythonPath += Seq(sparkHome, "python", "lib", "py4j-0.10.4-src.zip").mkString(File.separator)`
	`35`	`+ pythonPath += Seq(sparkHome, "python", "lib", "py4j-0.10.6-src.zip").mkString(File.separator)`
`36`	`36`	`}`
`37`	`37`	`pythonPath ++= SparkContext.jarOfObject(this)`
`38`	`38`	`pythonPath.mkString(File.pathSeparator)`
Original file line number	Diff line number	Diff line change
`@@ -29,4 +29,4 @@ The Python packaging for Spark is not intended to replace all of the other use c`
`29`	`29`
`30`	`30`	`## Python Requirements`
`31`	`31`
`32`		`-At its core PySpark depends on Py4J (currently version 0.10.4), but additional sub-packages have their own requirements (including numpy and pandas).`
	`32`	`+At its core PySpark depends on Py4J (currently version 0.10.6), but additional sub-packages have their own requirements (including numpy and pandas).`