[SPARK-13020][SQL][TEST] fix random generator for map type

when we generate map, we first randomly pick a length, then create a seq of key value pair with the expected length, and finally call `toMap`. However, `toMap` will remove all duplicated keys, which makes the actual map size much less than we expected.

This PR fixes this problem by put keys in a set first, to guarantee we have enough keys to build a map with expected length.

Author: Wenchen Fan <wenchen@databricks.com>

Closes #10930 from cloud-fan/random-generator.

2 files changed, 25 insertions, 4 deletions

diff --git a/sql/catalyst/src/test/scala/org/apache/spark/sql/RandomDataGenerator.scala b/sql/catalyst/src/test/scala/org/apache/spark/sql/RandomDataGenerator.scala
index 55efea80d1..7c173cbcee 100644
--- a/sql/catalyst/src/test/scala/org/apache/spark/sql/RandomDataGenerator.scala
+++ b/sql/catalyst/src/test/scala/org/apache/spark/sql/RandomDataGenerator.scala
@@ -47,9 +47,9 @@ object RandomDataGenerator {
    */
   private val PROBABILITY_OF_NULL: Float = 0.1f
 
-  private val MAX_STR_LEN: Int = 1024
-  private val MAX_ARR_SIZE: Int = 128
-  private val MAX_MAP_SIZE: Int = 128
+  final val MAX_STR_LEN: Int = 1024
+  final val MAX_ARR_SIZE: Int = 128
+  final val MAX_MAP_SIZE: Int = 128
 
   /**
    * Helper function for constructing a biased random number generator which returns "interesting"
@@ -208,7 +208,17 @@ object RandomDataGenerator {
             forType(valueType, nullable = valueContainsNull, rand)
         ) yield {
           () => {
-            Seq.fill(rand.nextInt(MAX_MAP_SIZE))((keyGenerator(), valueGenerator())).toMap
+            val length = rand.nextInt(MAX_MAP_SIZE)
+            val keys = scala.collection.mutable.HashSet(Seq.fill(length)(keyGenerator()): _*)
+            // In case the number of different keys is not enough, set a max iteration to avoid
+            // infinite loop.
+            var count = 0
+            while (keys.size < length && count < MAX_MAP_SIZE) {
+              keys += keyGenerator()
+              count += 1
+            }
+            val values = Seq.fill(keys.size)(valueGenerator())
+            keys.zip(values).toMap
           }
         }
       }
diff --git a/sql/catalyst/src/test/scala/org/apache/spark/sql/RandomDataGeneratorSuite.scala b/sql/catalyst/src/test/scala/org/apache/spark/sql/RandomDataGeneratorSuite.scala
index b8ccdf7516..9fba7924e9 100644
--- a/sql/catalyst/src/test/scala/org/apache/spark/sql/RandomDataGeneratorSuite.scala
+++ b/sql/catalyst/src/test/scala/org/apache/spark/sql/RandomDataGeneratorSuite.scala
@@ -95,4 +95,15 @@ class RandomDataGeneratorSuite extends SparkFunSuite {
     }
   }
 
+  test("check size of generated map") {
+    val mapType = MapType(IntegerType, IntegerType)
+    for (seed <- 1 to 1000) {
+      val generator = RandomDataGenerator.forType(
+        mapType, nullable = false, rand = new Random(seed)).get
+      val maps = Seq.fill(100)(generator().asInstanceOf[Map[Int, Int]])
+      val expectedTotalElements = 100 / 2 * RandomDataGenerator.MAX_MAP_SIZE
+      val deviation = math.abs(maps.map(_.size).sum - expectedTotalElements)
+      assert(deviation.toDouble / expectedTotalElements < 2e-1)
+    }
+  }
 }