[GitHub] [spark] allisonwang-db commented on a diff in pull request #42161: [SPARK-44479][PYTHON] Fix ArrowStreamPandasUDFSerializer to accept no-column pandas DataFrame

["allisonwang-db (via GitHub)" <gi...@apache.org>]

allisonwang-db commented on code in PR #42161:
URL: https://github.com/apache/spark/pull/42161#discussion_r1274354197


##########
python/pyspark/sql/pandas/serializers.py:
##########
@@ -385,37 +385,28 @@ def _create_struct_array(self, df, arrow_struct_type):
         """
         import pyarrow as pa
 
-        # Input partition and result pandas.DataFrame empty, make empty Arrays with struct
-        if len(df) == 0 and len(df.columns) == 0:
-            arrs_names = [
-                (pa.array([], type=field.type), field.name) for field in arrow_struct_type
-            ]
+        if len(df.columns) == 0:
+            return pa.array([{}] * len(df), arrow_struct_type)

Review Comment:
   Under what scenario would it be possible for a DataFrame to have rows but no columns? i.e len(df.columns) == 0 but len(df) > 0?



##########
python/pyspark/sql/tests/test_udtf.py:
##########
@@ -487,6 +487,14 @@ def eval(self, x: int):
 
         self.assertEqual(TestUDTF(lit(1)).collect(), [Row(x={1: "1"})])
 
+    def test_udtf_with_empty_output_types(self):
+        @udtf(returnType=StructType())
+        class TestUDTF:
+            def eval(self):
+                yield tuple()
+
+        self.assertEqual(TestUDTF().collect(), [Row()])

Review Comment:
   Let's use assertDataFrameEqual :)



##########
python/pyspark/sql/pandas/serializers.py:
##########
@@ -385,37 +385,28 @@ def _create_struct_array(self, df, arrow_struct_type):
         """
         import pyarrow as pa
 
-        # Input partition and result pandas.DataFrame empty, make empty Arrays with struct
-        if len(df) == 0 and len(df.columns) == 0:
-            arrs_names = [
-                (pa.array([], type=field.type), field.name) for field in arrow_struct_type
-            ]
+        if len(df.columns) == 0:
+            return pa.array([{}] * len(df), arrow_struct_type)
         # Assign result columns by schema name if user labeled with strings
         elif self._assign_cols_by_name and any(isinstance(name, str) for name in df.columns):

Review Comment:
   ```suggestion
           if self._assign_cols_by_name and any(isinstance(name, str) for name in df.columns):
   ```



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org