[SPARK-5865][API DOC] Add doc warnings for methods that return local data structures

rxin srowen
I work out note message for rdd.take function, please help to review.

If it's fine, I can apply to all other function later.

Author: Tommy YU <tummyyu@163.com>

Closes #10874 from Wenpei/spark-5865-add-warning-for-localdatastructure.

7 files changed, 72 insertions, 0 deletions

diff --git a/core/src/main/scala/org/apache/spark/api/java/JavaPairRDD.scala b/core/src/main/scala/org/apache/spark/api/java/JavaPairRDD.scala
index fb04472ee7..94d103588b 100644
--- a/core/src/main/scala/org/apache/spark/api/java/JavaPairRDD.scala
+++ b/core/src/main/scala/org/apache/spark/api/java/JavaPairRDD.scala
@@ -636,6 +636,9 @@ class JavaPairRDD[K, V](val rdd: RDD[(K, V)])
 
   /**
    * Return the key-value pairs in this RDD to the master as a Map.
+   *
+   * @note this method should only be used if the resulting data is expected to be small, as
+   * all the data is loaded into the driver's memory.
    */
   def collectAsMap(): java.util.Map[K, V] = mapAsSerializableJavaMap(rdd.collectAsMap())
 
diff --git a/core/src/main/scala/org/apache/spark/api/java/JavaRDDLike.scala b/core/src/main/scala/org/apache/spark/api/java/JavaRDDLike.scala
index 7340defabf..37c211fe70 100644
--- a/core/src/main/scala/org/apache/spark/api/java/JavaRDDLike.scala
+++ b/core/src/main/scala/org/apache/spark/api/java/JavaRDDLike.scala
@@ -327,6 +327,9 @@ trait JavaRDDLike[T, This <: JavaRDDLike[T, This]] extends Serializable {
 
   /**
    * Return an array that contains all of the elements in this RDD.
+   *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
    */
   def collect(): JList[T] =
     rdd.collect().toSeq.asJava
@@ -465,6 +468,9 @@ trait JavaRDDLike[T, This <: JavaRDDLike[T, This]] extends Serializable {
    * Take the first num elements of the RDD. This currently scans the partitions *one by one*, so
    * it will be slow if a lot of partitions are required. In that case, use collect() to get the
    * whole RDD instead.
+   *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
    */
   def take(num: Int): JList[T] =
     rdd.take(num).toSeq.asJava
@@ -548,6 +554,9 @@ trait JavaRDDLike[T, This <: JavaRDDLike[T, This]] extends Serializable {
   /**
    * Returns the top k (largest) elements from this RDD as defined by
    * the specified Comparator[T] and maintains the order.
+   *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
    * @param num k, the number of top elements to return
    * @param comp the comparator that defines the order
    * @return an array of top elements
@@ -559,6 +568,9 @@ trait JavaRDDLike[T, This <: JavaRDDLike[T, This]] extends Serializable {
   /**
    * Returns the top k (largest) elements from this RDD using the
    * natural ordering for T and maintains the order.
+   *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
    * @param num k, the number of top elements to return
    * @return an array of top elements
    */
@@ -570,6 +582,9 @@ trait JavaRDDLike[T, This <: JavaRDDLike[T, This]] extends Serializable {
   /**
    * Returns the first k (smallest) elements from this RDD as defined by
    * the specified Comparator[T] and maintains the order.
+   *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
    * @param num k, the number of elements to return
    * @param comp the comparator that defines the order
    * @return an array of top elements
@@ -601,6 +616,9 @@ trait JavaRDDLike[T, This <: JavaRDDLike[T, This]] extends Serializable {
   /**
    * Returns the first k (smallest) elements from this RDD using the
    * natural ordering for T while maintain the order.
+   *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
    * @param num k, the number of top elements to return
    * @return an array of top elements
    */
@@ -634,6 +652,9 @@ trait JavaRDDLike[T, This <: JavaRDDLike[T, This]] extends Serializable {
   /**
    * The asynchronous version of `collect`, which returns a future for
    * retrieving an array containing all of the elements in this RDD.
+   *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
    */
   def collectAsync(): JavaFutureAction[JList[T]] = {
     new JavaFutureActionWrapper(rdd.collectAsync(), (x: Seq[T]) => x.asJava)
@@ -642,6 +663,9 @@ trait JavaRDDLike[T, This <: JavaRDDLike[T, This]] extends Serializable {
   /**
    * The asynchronous version of the `take` action, which returns a
    * future for retrieving the first `num` elements of this RDD.
+   *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
    */
   def takeAsync(num: Int): JavaFutureAction[JList[T]] = {
     new JavaFutureActionWrapper(rdd.takeAsync(num), (x: Seq[T]) => x.asJava)
diff --git a/core/src/main/scala/org/apache/spark/rdd/PairRDDFunctions.scala b/core/src/main/scala/org/apache/spark/rdd/PairRDDFunctions.scala
index 33f2f0b44f..61905a8421 100644
--- a/core/src/main/scala/org/apache/spark/rdd/PairRDDFunctions.scala
+++ b/core/src/main/scala/org/apache/spark/rdd/PairRDDFunctions.scala
@@ -726,6 +726,9 @@ class PairRDDFunctions[K, V](self: RDD[(K, V)])
    *
    * Warning: this doesn't return a multimap (so if you have multiple values to the same key, only
    *          one value per key is preserved in the map returned)
+   *
+   * @note this method should only be used if the resulting data is expected to be small, as
+   * all the data is loaded into the driver's memory.
    */
   def collectAsMap(): Map[K, V] = self.withScope {
     val data = self.collect()
diff --git a/core/src/main/scala/org/apache/spark/rdd/RDD.scala b/core/src/main/scala/org/apache/spark/rdd/RDD.scala
index e8157cf4eb..a81a98b526 100644
--- a/core/src/main/scala/org/apache/spark/rdd/RDD.scala
+++ b/core/src/main/scala/org/apache/spark/rdd/RDD.scala
@@ -481,6 +481,9 @@ abstract class RDD[T: ClassTag](
   /**
    * Return a fixed-size sampled subset of this RDD in an array
    *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
+   *
    * @param withReplacement whether sampling is done with replacement
    * @param num size of the returned sample
    * @param seed seed for the random number generator
@@ -836,6 +839,9 @@ abstract class RDD[T: ClassTag](
 
   /**
    * Return an array that contains all of the elements in this RDD.
+   *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
    */
   def collect(): Array[T] = withScope {
     val results = sc.runJob(this, (iter: Iterator[T]) => iter.toArray)
@@ -1202,6 +1208,9 @@ abstract class RDD[T: ClassTag](
    * results from that partition to estimate the number of additional partitions needed to satisfy
    * the limit.
    *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
+   *
    * @note due to complications in the internal implementation, this method will raise
    * an exception if called on an RDD of `Nothing` or `Null`.
    */
@@ -1263,6 +1272,9 @@ abstract class RDD[T: ClassTag](
    *   // returns Array(6, 5)
    * }}}
    *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
+   *
    * @param num k, the number of top elements to return
    * @param ord the implicit ordering for T
    * @return an array of top elements
@@ -1283,6 +1295,9 @@ abstract class RDD[T: ClassTag](
    *   // returns Array(2, 3)
    * }}}
    *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
+   *
    * @param num k, the number of elements to return
    * @param ord the implicit ordering for T
    * @return an array of top elements
diff --git a/python/pyspark/rdd.py b/python/pyspark/rdd.py
index c285946254..fe2264a63c 100644
--- a/python/pyspark/rdd.py
+++ b/python/pyspark/rdd.py
@@ -426,6 +426,9 @@ class RDD(object):
         """
         Return a fixed-size sampled subset of this RDD.
 
+        Note that this method should only be used if the resulting array is expected
+        to be small, as all the data is loaded into the driver's memory.
+
         >>> rdd = sc.parallelize(range(0, 10))
         >>> len(rdd.takeSample(True, 20, 1))
         20
@@ -766,6 +769,8 @@ class RDD(object):
     def collect(self):
         """
         Return a list that contains all of the elements in this RDD.
+        Note that this method should only be used if the resulting array is expected
+        to be small, as all the data is loaded into the driver's memory.
         """
         with SCCallSiteSync(self.context) as css:
             port = self.ctx._jvm.PythonRDD.collectAndServe(self._jrdd.rdd())
@@ -1213,6 +1218,9 @@ class RDD(object):
         """
         Get the top N elements from a RDD.
 
+        Note that this method should only be used if the resulting array is expected
+        to be small, as all the data is loaded into the driver's memory.
+
         Note: It returns the list sorted in descending order.
 
         >>> sc.parallelize([10, 4, 2, 12, 3]).top(1)
@@ -1235,6 +1243,9 @@ class RDD(object):
         Get the N elements from a RDD ordered in ascending order or as
         specified by the optional key function.
 
+        Note that this method should only be used if the resulting array is expected
+        to be small, as all the data is loaded into the driver's memory.
+
         >>> sc.parallelize([10, 1, 2, 9, 3, 4, 5, 6, 7]).takeOrdered(6)
         [1, 2, 3, 4, 5, 6]
         >>> sc.parallelize([10, 1, 2, 9, 3, 4, 5, 6, 7], 2).takeOrdered(6, key=lambda x: -x)
@@ -1254,6 +1265,9 @@ class RDD(object):
         that partition to estimate the number of additional partitions needed
         to satisfy the limit.
 
+        Note that this method should only be used if the resulting array is expected
+        to be small, as all the data is loaded into the driver's memory.
+
         Translated from the Scala implementation in RDD#take().
 
         >>> sc.parallelize([2, 3, 4, 5, 6]).cache().take(2)
@@ -1511,6 +1525,9 @@ class RDD(object):
         """
         Return the key-value pairs in this RDD to the master as a dictionary.
 
+        Note that this method should only be used if the resulting data is expected
+        to be small, as all the data is loaded into the driver's memory.
+
         >>> m = sc.parallelize([(1, 2), (3, 4)]).collectAsMap()
         >>> m[1]
         2
diff --git a/python/pyspark/sql/dataframe.py b/python/pyspark/sql/dataframe.py
index 90a6b5d9c0..3a8c8305ee 100644
--- a/python/pyspark/sql/dataframe.py
+++ b/python/pyspark/sql/dataframe.py
@@ -739,6 +739,9 @@ class DataFrame(object):
     def head(self, n=None):
         """Returns the first ``n`` rows.
 
+        Note that this method should only be used if the resulting array is expected
+        to be small, as all the data is loaded into the driver's memory.
+
         :param n: int, default 1. Number of rows to return.
         :return: If n is greater than 1, return a list of :class:`Row`.
             If n is 1, return a single Row.
@@ -1330,6 +1333,9 @@ class DataFrame(object):
     def toPandas(self):
         """Returns the contents of this :class:`DataFrame` as Pandas ``pandas.DataFrame``.
 
+        Note that this method should only be used if the resulting Pandas's DataFrame is expected
+        to be small, as all the data is loaded into the driver's memory.
+
         This is only available if Pandas is installed and available.
 
         >>> df.toPandas()  # doctest: +SKIP
diff --git a/sql/core/src/main/scala/org/apache/spark/sql/DataFrame.scala b/sql/core/src/main/scala/org/apache/spark/sql/DataFrame.scala
index f15b926bd2..7aa08fb630 100644
--- a/sql/core/src/main/scala/org/apache/spark/sql/DataFrame.scala
+++ b/sql/core/src/main/scala/org/apache/spark/sql/DataFrame.scala
@@ -1384,6 +1384,10 @@ class DataFrame private[sql](
 
   /**
    * Returns the first `n` rows.
+   *
+   * @note this method should only be used if the resulting array is expected to be small, as
+   * all the data is loaded into the driver's memory.
+   *
    * @group action
    * @since 1.3.0
    */