[SPARK-6972][SQL] Add Coalesce to DataFrame

Author: Michael Armbrust <[email protected]>

Closes apache#5545 from marmbrus/addCoalesce and squashes the following commits:

9fdf3f6 [Michael Armbrust] [SPARK-6972][SQL] Add Coalesce to DataFrame

Author: marmbrus

sql/core/src/main/scala/org/apache/spark/sql/DataFrame.scala

@@ -908,6 +908,20 @@ class DataFrame private[sql](
       schema, needsConversion = false)
   }
 
+  /**
+   * Returns a new [[DataFrame]] that has exactly `numPartitions` partitions.
+   * Similar to coalesce defined on an [[RDD]], this operation results in a narrow dependency, e.g.
+   * if you go from 1000 partitions to 100 partitions, there will not be a shuffle, instead each of
+   * the 100 new partitions will claim 10 of the current partitions.
+   * @group rdd
+   */
+  override def coalesce(numPartitions: Int): DataFrame = {
+    sqlContext.createDataFrame(
+      queryExecution.toRdd.coalesce(numPartitions),
+      schema,
+      needsConversion = false)
+  }
+
   /**
    * Returns a new [[DataFrame]] that contains only the unique rows from this [[DataFrame]].
    * @group dfops

sql/core/src/main/scala/org/apache/spark/sql/RDDApi.scala

@@ -61,5 +61,7 @@ private[sql] trait RDDApi[T] {
 
   def repartition(numPartitions: Int): DataFrame
 
+  def coalesce(numPartitions: Int): DataFrame
+
   def distinct: DataFrame
 }

sql/core/src/test/scala/org/apache/spark/sql/DataFrameSuite.scala

@@ -178,6 +178,14 @@ class DataFrameSuite extends QueryTest {
       testData.select('key).collect().toSeq)
   }
 
+  test("coalesce") {
+    assert(testData.select('key).coalesce(1).rdd.partitions.size === 1)
+
+    checkAnswer(
+      testData.select('key).coalesce(1).select('key),
+      testData.select('key).collect().toSeq)
+  }
+
   test("groupBy") {
     checkAnswer(
       testData2.groupBy("a").agg($"a", sum($"b")),