address comments from @shivaram

Author: Davies Liu

R/pkg/R/RDD.R

@@ -238,7 +238,7 @@ setMethod("cache",
 #' @aliases persist,RDD-method
 setMethod("persist",
           signature(x = "RDD", newLevel = "character"),
-          function(x, newLevel) {
+          function(x, newLevel = "MEMORY_ONLY") {
             callJMethod(getJRDD(x), "persist", getStorageLevel(newLevel))
             x@env$isCached <- TRUE
             x

R/pkg/R/pairRDD.R

@@ -327,7 +327,7 @@ setMethod("reduceByKey",
               convertEnvsToList(keys, vals)
             }
             locallyReduced <- lapplyPartition(x, reduceVals)
-            shuffled <- partitionBy(locallyReduced, as.integer(numPartitions))
+            shuffled <- partitionBy(locallyReduced, numToInt(numPartitions))
             lapplyPartition(shuffled, reduceVals)
           })
 

docs/index.md

@@ -54,7 +54,7 @@ Example applications are also provided in Python. For example,
 
     ./bin/spark-submit examples/src/main/python/pi.py 10
 
-Spark also provides an experimental R API since 1.4 (only RDD and DtaFrame APIs included).
+Spark also provides an experimental R API since 1.4 (only RDD and DataFrames APIs included).
 To run Spark interactively in a R interpreter, use `bin/sparkR`:
 
     ./bin/sparkR --master local[2]

docs/programming-guide.md

@@ -867,7 +867,7 @@ There are three recommended ways to do this:
 For example, to pass a longer function, consider the code below:
 
 {% highlight r %}
-"""MyScript.py"""
+"""MyScript.R"""
 myFunc <- function(s) {
     words = strsplit(s, " ")[[1]]
     length(words)

docs/quick-start.md

@@ -225,7 +225,11 @@ For example, we'll define a `mymax` function to make this code easier to underst
 One common data flow pattern is MapReduce, as popularized by Hadoop. Spark can implement MapReduce flows easily:
 
 {% highlight r %}
-> wordCounts <- reduceByKey(map(flatMap(textFile, function(line) strsplit(line, " ")[[1]]), function(word) list(word, 1)), "+", 2)
+> wordCounts <- reduceByKey(
+    map(
+      flatMap(textFile, function(line) strsplit(line, " ")[[1]]),
+      function(word) list(word, 1)),
+    "+", 2)
 {% endhighlight %}
 
 Here, we combined the [`flatMap`](programming-guide.html#transformations), [`map`](programming-guide.html#transformations) and [`reduceByKey`](programming-guide.html#transformations) transformations to compute the per-word counts in the file as an RDD of (string, numeric) pairs. To collect the word counts in our shell, we can use the [`collect`](programming-guide.html#actions) action:
@@ -256,10 +260,10 @@ scala> linesWithSpark.cache()
 res7: spark.RDD[String] = spark.FilteredRDD@17e51082
 
 scala> linesWithSpark.count()
-res8: Long = 15
+res8: Long = 19
 
 scala> linesWithSpark.count()
-res9: Long = 15
+res9: Long = 19
 {% endhighlight %}
 
 It may seem silly to use Spark to explore and cache a 100-line text file. The interesting part is
@@ -274,10 +278,10 @@ a cluster, as described in the [programming guide](programming-guide.html#initia
 >>> linesWithSpark.cache()
 
 >>> linesWithSpark.count()
-15
+19
 
 >>> linesWithSpark.count()
-15
+19
 {% endhighlight %}
 
 It may seem silly to use Spark to explore and cache a 100-line text file. The interesting part is
@@ -292,10 +296,10 @@ a cluster, as described in the [programming guide](programming-guide.html#initia
 > cache(linesWithSpark)
 
 > count(linesWithSpark)
-[1] 15
+[1] 19
 
 > count(linesWithSpark)
-[1] 15
+[1] 19
 {% endhighlight %}
 
 It may seem silly to use Spark to explore and cache a 100-line text file. The interesting part is