SNOW-1952256: Add support for create_dataframe with FILE data type (#3322)

Author: sfc-gh-jdu

CHANGELOG.md

@@ -7,6 +7,7 @@
 #### Improvements
 
 - Invoking snowflake system procedures does not invoke an additional `describe procedure` call to check the return type of the procedure.
+- Added support for `Session.create_dataframe()` with the stage URL and FILE data type.
 
 #### Bug Fixes
 

docs/source/snowpark/types.rst

@@ -20,6 +20,7 @@ This package contains all Snowpark logical types.
     DateType
     DecimalType
     DoubleType
+    FileType
     FloatType
     Geography
     GeographyType

src/snowflake/snowpark/session.py

@@ -3549,6 +3549,7 @@ def convert_row_to_list(
                         TimestampType,
                         VariantType,
                         VectorType,
+                        FileType,
                     ),
                 )
                 else field.datatype
@@ -3675,7 +3676,6 @@ def convert_row_to_list(
                 project_columns.append(
                     parse_json(column(name)).cast(field.datatype).as_(name)
                 )
-            # TODO SNOW-1952256: Test file type in create_dataframe once it accepts full path
             elif isinstance(field.datatype, FileType):
                 project_columns.append(to_file(column(name)).as_(name))
             else:

tests/integ/test_dataframe.py

@@ -65,6 +65,7 @@
     when,
 )
 from snowflake.snowpark.types import (
+    FileType,
     ArrayType,
     BinaryType,
     BooleanType,
@@ -5180,3 +5181,30 @@ def test_create_dataframe_x_string_y_integer(session):
         Row(X="b", Y=2),
     ]
     assert result == expected_rows
+
+
+@pytest.mark.skipif(
+    "config.getoption('local_testing_mode', default=False)",
+    reason="File data type is not supported in Local Testing",
+)
+def test_create_dataframe_file_type(session, resources_path):
+    stage_name = Utils.random_name_for_temp_object(TempObjectType.STAGE)
+    session.sql(f"create or replace temp stage {stage_name}").collect()
+    test_files = TestFiles(resources_path)
+    session.file.put(
+        test_files.test_file_csv, f"@{stage_name}", auto_compress=False, overwrite=True
+    )
+    session.file.put(
+        test_files.test_dog_image, f"@{stage_name}", auto_compress=False, overwrite=True
+    )
+    csv_url = f"@{stage_name}/testCSV.csv"
+    image_url = f"@{stage_name}/dog.jpg"
+    df = session.create_dataframe(
+        [csv_url, image_url],
+        schema=StructType([StructField("col", FileType(), nullable=True)]),
+    )
+    result = df.collect()
+    csv_row = json.loads(result[0][0])
+    assert csv_row["CONTENT_TYPE"] == "text/csv"
+    image_row = json.loads(result[1][0])
+    assert image_row["CONTENT_TYPE"] == "image/jpeg"

tests/utils.py

@@ -1554,6 +1554,10 @@ def test_xxe_xml(self):
     def test_nested_xml(self):
         return os.path.join(self.resources_path, "nested.xml")
 
+    @property
+    def test_dog_image(self):
+        return os.path.join(self.resources_path, "dog.jpg")
+
 
 class TypeMap(NamedTuple):
     col_name: str