Call cross join path in PySpark join rather than throwing NPE

Author: HyukjinKwon

python/pyspark/sql/dataframe.py

@@ -833,6 +833,8 @@ def join(self, other, on=None, how=None):
         else:
             if how is None:
                 how = "inner"
+            if on is None:
+                on = self._jseq([])
             assert isinstance(how, basestring), "how should be basestring"
             jdf = self._jdf.join(other._jdf, on, how)
         return DataFrame(jdf, self.sql_ctx)

python/pyspark/sql/tests.py

@@ -2021,6 +2021,11 @@ def test_toDF_with_schema_string(self):
         self.assertEqual(df.schema.simpleString(), "struct<value:int>")
         self.assertEqual(df.collect(), [Row(key=i) for i in range(100)])
 
+    def test_join_without_on(self):
+        self.assertRaises(
+            AnalysisException,
+            lambda: self.spark.range(1).join(self.spark.range(1), how="inner").collect())
+
     # Regression test for invalid join methods when on is None, Spark-14761
     def test_invalid_join_method(self):
         df1 = self.spark.createDataFrame([("Alice", 5), ("Bob", 8)], ["name", "age"])