Add regression test.

Author: JoshRosen

python/pyspark/sql/tests.py

@@ -1854,6 +1854,24 @@ def test_collect_functions(self):
             sorted(df.select(functions.collect_list(df.value).alias('r')).collect()[0].r),
             ["1", "2", "2", "2"])
 
+    def test_limit_and_take(self):
+        df = self.spark.range(1, 1000, numPartitions=10)
+
+        def assert_runs_only_one_job_stage_and_task(job_group_name, f):
+            tracker = self.sc.statusTracker()
+            self.sc.setJobGroup(job_group_name, description="")
+            f()
+            jobs = tracker.getJobIdsForGroup(job_group_name)
+            self.assertEqual(1, len(jobs))
+            stages = tracker.getJobInfo(jobs[0]).stageIds
+            self.assertEqual(1, len(stages))
+            self.assertEqual(1, tracker.getStageInfo(stages[0]).numTasks)
+
+        # Regression test for SPARK-10731: take should delegate to Scala implementation
+        assert_runs_only_one_job_stage_and_task("take", lambda: df.take(1))
+        # Regression test for SPARK-17514: limit(n).collect() should the perform same as take(n)
+        assert_runs_only_one_job_stage_and_task("collect_limit", lambda: df.limit(1).collect())
+
 
 if __name__ == "__main__":
     from pyspark.sql.tests import *