add numRows param in DataFrame.show()

python/pyspark/sql/dataframe.py

@@ -272,9 +272,9 @@ def isLocal(self):
         """
         return self._jdf.isLocal()
 
-    def show(self):
+    def show(self, numRows=20):
         """
-        Print the first 20 rows.
+        Print the first n rows.
 
         >>> df
         DataFrame[age: int, name: string]
@@ -283,7 +283,7 @@ def show(self):
         2   Alice
         5   Bob
         """
-        print self._jdf.showString().encode('utf8', 'ignore')
+        print self._jdf.showString(numRows).encode('utf8', 'ignore')
 
     def __repr__(self):
         return "DataFrame[%s]" % (", ".join("%s: %s" % c for c in self.dtypes))

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

@@ -159,9 +159,11 @@ class DataFrame protected[sql](
 
   /**
    * Internal API for Python
+   * @param numRows Number of rows to show
    */
-  private[sql] def showString(): String = {
-    val data = take(20)
+  private[sql] def showString(numRows: Int = 20): String = {
+    val size = count()
+    val data = if (numRows > size) collect() else take(numRows)
     val numCols = schema.fieldNames.length
 
     // For cells that are beyond 20 characters, replace it with the first 17 and "..."
@@ -293,9 +295,10 @@ class DataFrame protected[sql](
    *   1983  03    0.410516        0.442194
    *   1984  04    0.450090        0.483521
    * }}}
+   * @param numRows Number of rows to show
    * @group basic
    */
-  def show(): Unit = println(showString())
+  def show(numRows: Int = 20): Unit = println(showString(numRows))
 
   /**
    * Cartesian join with another [[DataFrame]].