aboutsummaryrefslogtreecommitdiff
path: root/python/pyspark/tests.py
diff options
context:
space:
mode:
Diffstat (limited to 'python/pyspark/tests.py')
-rw-r--r--python/pyspark/tests.py28
1 files changed, 28 insertions, 0 deletions
diff --git a/python/pyspark/tests.py b/python/pyspark/tests.py
index 2e7c2750a8..b687d695b0 100644
--- a/python/pyspark/tests.py
+++ b/python/pyspark/tests.py
@@ -607,6 +607,34 @@ class TestSQL(PySparkTestCase):
[res] = self.sqlCtx.sql("SELECT MYUDF('')").collect()
self.assertEqual("", res[0])
+ def test_basic_functions(self):
+ rdd = self.sc.parallelize(['{"foo":"bar"}', '{"foo":"baz"}'])
+ srdd = self.sqlCtx.jsonRDD(rdd)
+ srdd.count()
+ srdd.collect()
+ srdd.schemaString()
+ srdd.schema()
+
+ # cache and checkpoint
+ self.assertFalse(srdd.is_cached)
+ srdd.persist()
+ srdd.unpersist()
+ srdd.cache()
+ self.assertTrue(srdd.is_cached)
+ self.assertFalse(srdd.isCheckpointed())
+ self.assertEqual(None, srdd.getCheckpointFile())
+
+ srdd = srdd.coalesce(2, True)
+ srdd = srdd.repartition(3)
+ srdd = srdd.distinct()
+ srdd.intersection(srdd)
+ self.assertEqual(2, srdd.count())
+
+ srdd.registerTempTable("temp")
+ srdd = self.sqlCtx.sql("select foo from temp")
+ srdd.count()
+ srdd.collect()
+
class TestIO(PySparkTestCase):