SNOW-1803811: Allow mixed-case field names for struct type columns

CHANGELOG.md

@@ -21,6 +21,7 @@
 - Added support for applying Snowpark Python function `snowflake_cortex_sentiment`.
 - Added support for `DataFrame.map`.
 - Added support for `DataFrame.from_dict` and `DataFrame.from_records`.
+- Added support for mixed case field names in struct type columns.
 
 #### Improvements
 - Improve performance of `DataFrame.map`, `Series.apply` and `Series.map` methods by mapping numpy functions to snowpark functions if possible.

src/snowflake/snowpark/_internal/type_utils.py

@@ -30,6 +30,7 @@
     get_origin,
 )
 
+import snowflake.snowpark.context as context
 import snowflake.snowpark.types  # type: ignore
 from snowflake.connector.constants import FIELD_ID_TO_NAME
 from snowflake.connector.cursor import ResultMetadata
@@ -157,9 +158,12 @@ def convert_metadata_to_sp_type(
             return StructType(
                 [
                     StructField(
-                        quote_name(field.name, keep_case=True),
+                        field.name
+                        if context._should_use_structured_type_semantics
+                        else quote_name(field.name, keep_case=True),
                         convert_metadata_to_sp_type(field, max_string_size),
                         nullable=field.is_nullable,
+                        _is_column=False,
                     )
                     for field in metadata.fields
                 ],

src/snowflake/snowpark/column.py

@@ -91,6 +91,9 @@
     StringType,
     TimestampTimeZone,
     TimestampType,
+    ArrayType,
+    MapType,
+    StructType,
 )
 from snowflake.snowpark.window import Window, WindowSpec
 
@@ -917,6 +920,9 @@ def _cast(
         if isinstance(to, str):
             to = type_string_to_type_object(to)
 
+        if isinstance(to, (ArrayType, MapType, StructType)):
+            to = to._as_nested()
+
         if self._ast is None:
             _emit_ast = False
 

src/snowflake/snowpark/context.py

@@ -21,6 +21,10 @@
 _should_continue_registration: Optional[Callable[..., bool]] = None
 
 
+# Global flag that determines if structured type semantics should be used
+_should_use_structured_type_semantics = False
+
+
 def get_active_session() -> "snowflake.snowpark.Session":
     """Returns the current active Snowpark session.
 

src/snowflake/snowpark/types.py

@@ -16,6 +16,7 @@
 
 # Use correct version from here:
 from snowflake.snowpark._internal.utils import installed_pandas, pandas, quote_name
+import snowflake.snowpark.context as context
 
 # TODO: connector installed_pandas is broken. If pyarrow is not installed, but pandas is this function returns the wrong answer.
 # The core issue is that in the connector detection of both pandas/arrow are mixed, which is wrong.
@@ -341,6 +342,14 @@ def __init__(
     def __repr__(self) -> str:
         return f"ArrayType({repr(self.element_type) if self.element_type else ''})"
 
+    def _as_nested(self) -> "ArrayType":
+        if not context._should_use_structured_type_semantics:
+            return self
+        element_type = self.element_type
+        if isinstance(element_type, (ArrayType, MapType, StructType)):
+            element_type = element_type._as_nested()
+        return ArrayType(element_type, self.structured)
+
     def is_primitive(self):
         return False
 
@@ -391,6 +400,14 @@ def __repr__(self) -> str:
     def is_primitive(self):
         return False
 
+    def _as_nested(self) -> "MapType":
+        if not context._should_use_structured_type_semantics:
+            return self
+        value_type = self.value_type
+        if isinstance(value_type, (ArrayType, MapType, StructType)):
+            value_type = value_type._as_nested()
+        return MapType(self.key_type, value_type, self.structured)
+
     @classmethod
     def from_json(cls, json_dict: Dict[str, Any]) -> "MapType":
         return MapType(
@@ -552,29 +569,46 @@ def __init__(
         column_identifier: Union[ColumnIdentifier, str],
         datatype: DataType,
         nullable: bool = True,
+        _is_column: bool = True,
     ) -> None:
-        self.column_identifier = (
-            ColumnIdentifier(column_identifier)
-            if isinstance(column_identifier, str)
-            else column_identifier
-        )
+        self.name = column_identifier
+        self._is_column = _is_column
         self.datatype = datatype
         self.nullable = nullable
 
     @property
     def name(self) -> str:
-        """Returns the column name."""
-        return self.column_identifier.name
+        if self._is_column or not context._should_use_structured_type_semantics:
+            return self.column_identifier.name
+        else:
+            return self._name
 
     @name.setter
-    def name(self, n: str) -> None:
-        self.column_identifier = ColumnIdentifier(n)
+    def name(self, n: Union[ColumnIdentifier, str]) -> None:
+        if isinstance(n, ColumnIdentifier):
+            self._name = n.name
+            self.column_identifier = n
+        else:
+            self._name = n
+            self.column_identifier = ColumnIdentifier(n)
+
+    def _as_nested(self) -> "StructField":
+        if not context._should_use_structured_type_semantics:
+            return self
+        datatype = self.datatype
+        if isinstance(datatype, (ArrayType, MapType, StructType)):
+            datatype = datatype._as_nested()
+        # Nested StructFields do not follow column naming conventions
+        return StructField(self._name, datatype, self.nullable, _is_column=False)
 
     def __repr__(self) -> str:
         return f"StructField({self.name!r}, {repr(self.datatype)}, nullable={self.nullable})"
 
     def __eq__(self, other):
-        return isinstance(other, self.__class__) and self.__dict__ == other.__dict__
+        return isinstance(other, self.__class__) and (
+            (self.name, self._is_column, self.datatype, self.nullable)
+            == (other.name, other._is_column, other.datatype, other.nullable)
+        )
 
     @classmethod
     def from_json(cls, json_dict: Dict[str, Any]) -> "StructField":
@@ -620,30 +654,41 @@ def __init__(
         self, fields: Optional[List["StructField"]] = None, structured=False
     ) -> None:
         self.structured = structured
-        if fields is None:
-            fields = []
-        self.fields = fields
+        self.fields = []
+        for field in fields or []:
+            self.add(field)
 
     def add(
         self,
         field: Union[str, ColumnIdentifier, "StructField"],
         datatype: Optional[DataType] = None,
         nullable: Optional[bool] = True,
     ) -> "StructType":
-        if isinstance(field, StructField):
-            self.fields.append(field)
-        elif isinstance(field, (str, ColumnIdentifier)):
+        if isinstance(field, (str, ColumnIdentifier)):
             if datatype is None:
                 raise ValueError(
                     "When field argument is str or ColumnIdentifier, datatype must not be None."
                 )
-            self.fields.append(StructField(field, datatype, nullable))
-        else:
+            field = StructField(field, datatype, nullable)
+        elif not isinstance(field, StructField):
             raise ValueError(
                 f"field argument must be one of str, ColumnIdentifier or StructField. Got: '{type(field)}'"
             )
+
+        # Nested data does not follow the same schema conventions as top level fields.
+        if isinstance(field.datatype, (ArrayType, MapType, StructType)):
+            field.datatype = field.datatype._as_nested()
+
+        self.fields.append(field)
         return self
 
+    def _as_nested(self) -> "StructType":
+        if not context._should_use_structured_type_semantics:
+            return self
+        return StructType(
+            [field._as_nested() for field in self.fields], self.structured
+        )
+
     @classmethod
     def _from_attributes(cls, attributes: list) -> "StructType":
         return cls([StructField(a.name, a.datatype, a.nullable) for a in attributes])