Merge branch 'master' of github.com:apache/spark into r-submit

Author: brkyvz

R/README.md

@@ -6,7 +6,7 @@ SparkR is an R package that provides a light-weight frontend to use Spark from R
 
 #### Build Spark
 
-Build Spark with [Maven](http://spark.apache.org/docs/latest/building-spark.html#building-with-buildmvn) and include the `-PsparkR` profile to build the R package. For example to use the default Hadoop versions you can run
+Build Spark with [Maven](http://spark.apache.org/docs/latest/building-spark.html#building-with-buildmvn) and include the `-Psparkr` profile to build the R package. For example to use the default Hadoop versions you can run
 ```
   build/mvn -DskipTests -Psparkr package
 ```

R/pkg/R/generics.R

@@ -20,7 +20,8 @@
 # @rdname aggregateRDD
 # @seealso reduce
 # @export
-setGeneric("aggregateRDD", function(x, zeroValue, seqOp, combOp) { standardGeneric("aggregateRDD") })
+setGeneric("aggregateRDD",
+           function(x, zeroValue, seqOp, combOp) { standardGeneric("aggregateRDD") })
 
 # @rdname cache-methods
 # @export

R/pkg/R/pairRDD.R

@@ -560,8 +560,8 @@ setMethod("join",
 # Left outer join two RDDs
 #
 # @description
-# \code{leftouterjoin} This function left-outer-joins two RDDs where every element is of the form list(K, V).
-# The key types of the two RDDs should be the same.
+# \code{leftouterjoin} This function left-outer-joins two RDDs where every element is of
+# the form list(K, V). The key types of the two RDDs should be the same.
 #
 # @param x An RDD to be joined. Should be an RDD where each element is
 #             list(K, V).
@@ -597,8 +597,8 @@ setMethod("leftOuterJoin",
 # Right outer join two RDDs
 #
 # @description
-# \code{rightouterjoin} This function right-outer-joins two RDDs where every element is of the form list(K, V).
-# The key types of the two RDDs should be the same.
+# \code{rightouterjoin} This function right-outer-joins two RDDs where every element is of
+# the form list(K, V). The key types of the two RDDs should be the same.
 #
 # @param x An RDD to be joined. Should be an RDD where each element is
 #             list(K, V).
@@ -634,8 +634,8 @@ setMethod("rightOuterJoin",
 # Full outer join two RDDs
 #
 # @description
-# \code{fullouterjoin} This function full-outer-joins two RDDs where every element is of the form list(K, V).
-# The key types of the two RDDs should be the same.
+# \code{fullouterjoin} This function full-outer-joins two RDDs where every element is of
+# the form list(K, V). The key types of the two RDDs should be the same.
 #
 # @param x An RDD to be joined. Should be an RDD where each element is
 #             list(K, V).

R/pkg/R/sparkR.R

@@ -105,11 +105,12 @@ sparkR.init <- function(
   sparkPackages = "") {
 
   if (exists(".sparkRjsc", envir = .sparkREnv)) {
-    cat("Re-using existing Spark Context. Please stop SparkR with sparkR.stop() or restart R to create a new Spark Context\n")
+    cat(paste("Re-using existing Spark Context.",
+              "Please stop SparkR with sparkR.stop() or restart R to create a new Spark Context\n"))
     return(get(".sparkRjsc", envir = .sparkREnv))
   }
 
-  sparkMem <- Sys.getenv("SPARK_MEM", "512m")
+  sparkMem <- Sys.getenv("SPARK_MEM", "1024m")
   jars <- suppressWarnings(normalizePath(as.character(sparkJars)))
 
   # Classpath separator is ";" on Windows
@@ -180,14 +181,16 @@ sparkR.init <- function(
 
   sparkExecutorEnvMap <- new.env()
   if (!any(names(sparkExecutorEnv) == "LD_LIBRARY_PATH")) {
-    sparkExecutorEnvMap[["LD_LIBRARY_PATH"]] <- paste0("$LD_LIBRARY_PATH:",Sys.getenv("LD_LIBRARY_PATH"))
+    sparkExecutorEnvMap[["LD_LIBRARY_PATH"]] <-
+      paste0("$LD_LIBRARY_PATH:",Sys.getenv("LD_LIBRARY_PATH"))
   }
   for (varname in names(sparkExecutorEnv)) {
     sparkExecutorEnvMap[[varname]] <- sparkExecutorEnv[[varname]]
   }
 
   nonEmptyJars <- Filter(function(x) { x != "" }, jars)
-  localJarPaths <- sapply(nonEmptyJars, function(j) { utils::URLencode(paste("file:", uriSep, j, sep = "")) })
+  localJarPaths <- sapply(nonEmptyJars,
+                          function(j) { utils::URLencode(paste("file:", uriSep, j, sep = "")) })
 
   # Set the start time to identify jobjs
   # Seconds resolution is good enough for this purpose, so use ints

R/pkg/R/utils.R

@@ -334,18 +334,21 @@ getStorageLevel <- function(newLevel = c("DISK_ONLY",
                                          "MEMORY_ONLY_SER_2",
                                          "OFF_HEAP")) {
   match.arg(newLevel)
+  storageLevelClass <- "org.apache.spark.storage.StorageLevel"
   storageLevel <- switch(newLevel,
-                         "DISK_ONLY" = callJStatic("org.apache.spark.storage.StorageLevel", "DISK_ONLY"),
-                         "DISK_ONLY_2" = callJStatic("org.apache.spark.storage.StorageLevel", "DISK_ONLY_2"),
-                         "MEMORY_AND_DISK" = callJStatic("org.apache.spark.storage.StorageLevel", "MEMORY_AND_DISK"),
-                         "MEMORY_AND_DISK_2" = callJStatic("org.apache.spark.storage.StorageLevel", "MEMORY_AND_DISK_2"),
-                         "MEMORY_AND_DISK_SER" = callJStatic("org.apache.spark.storage.StorageLevel", "MEMORY_AND_DISK_SER"),
-                         "MEMORY_AND_DISK_SER_2" = callJStatic("org.apache.spark.storage.StorageLevel", "MEMORY_AND_DISK_SER_2"),
-                         "MEMORY_ONLY" = callJStatic("org.apache.spark.storage.StorageLevel", "MEMORY_ONLY"),
-                         "MEMORY_ONLY_2" = callJStatic("org.apache.spark.storage.StorageLevel", "MEMORY_ONLY_2"),
-                         "MEMORY_ONLY_SER" = callJStatic("org.apache.spark.storage.StorageLevel", "MEMORY_ONLY_SER"),
-                         "MEMORY_ONLY_SER_2" = callJStatic("org.apache.spark.storage.StorageLevel", "MEMORY_ONLY_SER_2"),
-                         "OFF_HEAP" = callJStatic("org.apache.spark.storage.StorageLevel", "OFF_HEAP"))
+                         "DISK_ONLY" = callJStatic(storageLevelClass, "DISK_ONLY"),
+                         "DISK_ONLY_2" = callJStatic(storageLevelClass, "DISK_ONLY_2"),
+                         "MEMORY_AND_DISK" = callJStatic(storageLevelClass, "MEMORY_AND_DISK"),
+                         "MEMORY_AND_DISK_2" = callJStatic(storageLevelClass, "MEMORY_AND_DISK_2"),
+                         "MEMORY_AND_DISK_SER" = callJStatic(storageLevelClass,
+                                                             "MEMORY_AND_DISK_SER"),
+                         "MEMORY_AND_DISK_SER_2" = callJStatic(storageLevelClass,
+                                                               "MEMORY_AND_DISK_SER_2"),
+                         "MEMORY_ONLY" = callJStatic(storageLevelClass, "MEMORY_ONLY"),
+                         "MEMORY_ONLY_2" = callJStatic(storageLevelClass, "MEMORY_ONLY_2"),
+                         "MEMORY_ONLY_SER" = callJStatic(storageLevelClass, "MEMORY_ONLY_SER"),
+                         "MEMORY_ONLY_SER_2" = callJStatic(storageLevelClass, "MEMORY_ONLY_SER_2"),
+                         "OFF_HEAP" = callJStatic(storageLevelClass, "OFF_HEAP"))
 }
 
 # Utility function for functions where an argument needs to be integer but we want to allow
@@ -545,9 +548,11 @@ mergePartitions <- function(rdd, zip) {
         lengthOfKeys <- part[[len - lengthOfValues]]
         stopifnot(len == lengthOfKeys + lengthOfValues)
 
-        # For zip operation, check if corresponding partitions of both RDDs have the same number of elements.
+        # For zip operation, check if corresponding partitions
+        # of both RDDs have the same number of elements.
         if (zip && lengthOfKeys != lengthOfValues) {
-          stop("Can only zip RDDs with same number of elements in each pair of corresponding partitions.")
+          stop(paste("Can only zip RDDs with same number of elements",
+                     "in each pair of corresponding partitions."))
         }
 
         if (lengthOfKeys > 1) {

R/pkg/inst/tests/test_binaryFile.R

@@ -82,7 +82,7 @@ test_that("saveAsObjectFile()/objectFile() works with multiple paths", {
   saveAsObjectFile(rdd2, fileName2)
 
   rdd <- objectFile(sc, c(fileName1, fileName2))
-  expect_true(count(rdd) == 2)
+  expect_equal(count(rdd), 2)
 
   unlink(fileName1, recursive = TRUE)
   unlink(fileName2, recursive = TRUE)

R/pkg/inst/tests/test_binary_function.R

@@ -38,13 +38,13 @@ test_that("union on two RDDs", {
   union.rdd <- unionRDD(rdd, text.rdd)
   actual <- collect(union.rdd)
   expect_equal(actual, c(as.list(nums), mockFile))
-  expect_true(getSerializedMode(union.rdd) == "byte")
+  expect_equal(getSerializedMode(union.rdd), "byte")
 
   rdd<- map(text.rdd, function(x) {x})
   union.rdd <- unionRDD(rdd, text.rdd)
   actual <- collect(union.rdd)
   expect_equal(actual, as.list(c(mockFile, mockFile)))
-  expect_true(getSerializedMode(union.rdd) == "byte")
+  expect_equal(getSerializedMode(union.rdd), "byte")
 
   unlink(fileName)
 })

R/pkg/inst/tests/test_includeJAR.R

@@ -18,8 +18,8 @@ context("include an external JAR in SparkContext")
 
 runScript <- function() {
   sparkHome <- Sys.getenv("SPARK_HOME")
-  jarPath <- paste("--jars",
-                   shQuote(file.path(sparkHome, "R/lib/SparkR/test_support/sparktestjar_2.10-1.0.jar")))
+  sparkTestJarPath <- "R/lib/SparkR/test_support/sparktestjar_2.10-1.0.jar"
+  jarPath <- paste("--jars", shQuote(file.path(sparkHome, sparkTestJarPath)))
   scriptPath <- file.path(sparkHome, "R/lib/SparkR/tests/jarTest.R")
   submitPath <- file.path(sparkHome, "bin/spark-submit")
   res <- system2(command = submitPath,
@@ -31,7 +31,7 @@ runScript <- function() {
 test_that("sparkJars tag in SparkContext", {
   testOutput <- runScript()
   helloTest <- testOutput[1]
-  expect_true(helloTest == "Hello, Dave")
+  expect_equal(helloTest, "Hello, Dave")
   basicFunction <- testOutput[2]
-  expect_true(basicFunction == 4L)
+  expect_equal(basicFunction, "4")
 })

R/pkg/inst/tests/test_parallelize_collect.R

@@ -57,7 +57,7 @@ test_that("parallelize() on simple vectors and lists returns an RDD", {
              strListRDD2)
 
   for (rdd in rdds) {
-    expect_true(inherits(rdd, "RDD"))
+    expect_is(rdd, "RDD")
     expect_true(.hasSlot(rdd, "jrdd")
                 && inherits(rdd@jrdd, "jobj")
                 && isInstanceOf(rdd@jrdd, "org.apache.spark.api.java.JavaRDD"))

R/pkg/inst/tests/test_rdd.R

@@ -33,9 +33,9 @@ test_that("get number of partitions in RDD", {
 })
 
 test_that("first on RDD", {
-  expect_true(first(rdd) == 1)
+  expect_equal(first(rdd), 1)
   newrdd <- lapply(rdd, function(x) x + 1)
-  expect_true(first(newrdd) == 2)
+  expect_equal(first(newrdd), 2)
 })
 
 test_that("count and length on RDD", {
@@ -669,27 +669,31 @@ test_that("fullOuterJoin() on pairwise RDDs", {
   rdd1 <- parallelize(sc, list(list(1,2), list(1,3), list(3,3)))
   rdd2 <- parallelize(sc, list(list(1,1), list(2,4)))
   actual <- collect(fullOuterJoin(rdd1, rdd2, 2L))
-  expected <- list(list(1, list(2, 1)), list(1, list(3, 1)), list(2, list(NULL, 4)), list(3, list(3, NULL)))
+  expected <- list(list(1, list(2, 1)), list(1, list(3, 1)),
+                   list(2, list(NULL, 4)), list(3, list(3, NULL)))
   expect_equal(sortKeyValueList(actual), sortKeyValueList(expected))
 
   rdd1 <- parallelize(sc, list(list("a",2), list("a",3), list("c", 1)))
   rdd2 <- parallelize(sc, list(list("a",1), list("b",4)))
   actual <- collect(fullOuterJoin(rdd1, rdd2, 2L))
-  expected <- list(list("b", list(NULL, 4)), list("a", list(2, 1)), list("a", list(3, 1)), list("c", list(1, NULL)))
+  expected <- list(list("b", list(NULL, 4)), list("a", list(2, 1)),
+                   list("a", list(3, 1)), list("c", list(1, NULL)))
   expect_equal(sortKeyValueList(actual),
                sortKeyValueList(expected))
 
   rdd1 <- parallelize(sc, list(list(1,1), list(2,2)))
   rdd2 <- parallelize(sc, list(list(3,3), list(4,4)))
   actual <- collect(fullOuterJoin(rdd1, rdd2, 2L))
   expect_equal(sortKeyValueList(actual),
-               sortKeyValueList(list(list(1, list(1, NULL)), list(2, list(2, NULL)), list(3, list(NULL, 3)), list(4, list(NULL, 4)))))
+               sortKeyValueList(list(list(1, list(1, NULL)), list(2, list(2, NULL)),
+                                     list(3, list(NULL, 3)), list(4, list(NULL, 4)))))
 
   rdd1 <- parallelize(sc, list(list("a",1), list("b",2)))
   rdd2 <- parallelize(sc, list(list("c",3), list("d",4)))
   actual <- collect(fullOuterJoin(rdd1, rdd2, 2L))
   expect_equal(sortKeyValueList(actual),
-               sortKeyValueList(list(list("a", list(1, NULL)), list("b", list(2, NULL)), list("d", list(NULL, 4)), list("c", list(NULL, 3)))))
+               sortKeyValueList(list(list("a", list(1, NULL)), list("b", list(2, NULL)),
+                                     list("d", list(NULL, 4)), list("c", list(NULL, 3)))))
 })
 
 test_that("sortByKey() on pairwise RDDs", {