Fix load parameter *unknow* propagation

When deserializing a data structure with the load method, the *unknown* was not propagated to the loading of nested data structures. As result, if a unknown field was present into a nested data structure a ValidationError was raised even if the load methd was called with *unknown=EXCLUDE*. This commit ensures that this parameter is now propagated also to the loading of nested data structures.
fixes #1428

Author: lmignon

AUTHORS.rst

@@ -140,3 +140,4 @@ Contributors (chronological)
 - `@phrfpeixoto <https://github.com/phrfpeixoto>`_
 - `@jceresini <https://github.com/jceresini>`_
 - Nikolay Shebanov `@killthekitten <https://github.com/killthekitten>`_
+- Laurent Mignon `@lmignon <https://github.com/lmignon>`_ 

docs/quickstart.rst

@@ -427,6 +427,55 @@ or when calling `load <Schema.load>`.
 
 The ``unknown`` option value set in :meth:`load <Schema.load>` will override the value applied at instantiation time, which itself will override the value defined in the *class Meta*.
 
+The ``unknown`` option value set in :meth:`load <Schema.load>` is propagated to nested Schema. It's not the case when the value is applied at instantiation time nor when the value is defined in the *class Meta*.
+
+.. code-block:: python
+
+    from marshmallow import fields, Schema, EXCLUDE, INCLUDE, RAISE
+
+
+    class UserSchema(Schema):
+        class Meta:
+            unknown = RAISE
+
+        name = fields.String()
+
+
+    class BlogSchema(Schema):
+        class Meta:
+            unknown = EXCLUDE
+
+        title = fields.String()
+        author = fields.Nested(UserSchema)
+
+
+    BlogSchema().load(
+        {
+            "title": "Some title",
+            "unknown_field": "value",
+            "author": {"name": "Author name", "unknown_field": "value"},
+        }
+    )  # => ValidationError: {'author': {'unknown_field': ['Unknown field.']}}
+
+    BlogSchema().load(
+        {
+            "title": "Some title",
+            "unknown_field": "value",
+            "author": {"name": "Author name", "unknown_field": "value"},
+        },
+        unknown=EXCLUDE,
+    )  # => {'author': {'name': 'Author name'}, 'title': 'Some title'}
+
+
+    BogSchema.load(
+        {
+            "title": "Some title",
+            "unknown_field": "value",
+            "author": {"name": "Author name", "unknown_field": "value"},
+        },
+        unknown=INCLUDE,
+    )  # => {'author': {'name': 'Author name', 'unknown_field': 'value'}, 'title': 'Some title', 'unknown_field': 'value'}
+
 This order of precedence allows you to change the behavior of a schema for different contexts.
 
 

src/marshmallow/fields.py

@@ -567,29 +567,36 @@ def _test_collection(self, value, many=False):
         if many and not utils.is_collection(value):
             raise self.make_error("type", input=value, type=value.__class__.__name__)
 
-    def _load(self, value, data, partial=None, many=False):
+    def _load(self, value, data, partial=None, many=False, unknown=None):
         many = self.schema.many or self.many or many
+        unknown = unknown or self.unknown
         try:
             valid_data = self.schema.load(
-                value, unknown=self.unknown, partial=partial, many=many
+                value, unknown=unknown, partial=partial, many=many
             )
         except ValidationError as error:
             raise ValidationError(
                 error.messages, valid_data=error.valid_data
             ) from error
         return valid_data
 
-    def _deserialize(self, value, attr, data, partial=None, many=False, **kwargs):
+    def _deserialize(
+        self, value, attr, data, partial=None, many=False, unknown=None, **kwargs
+    ):
         """Same as :meth:`Field._deserialize` with additional ``partial`` argument.
 
         :param bool|tuple partial: For nested schemas, the ``partial``
             parameter passed to `Schema.load`.
+        :param unknown: For nested schemas, the ``unknown``
+            parameter passed to `Schema.load`..
 
         .. versionchanged:: 3.0.0
             Add ``partial`` parameter.
+        .. versionchanged:: 3.2.2
+            Add ``unknown`` parameter.
         """
         self._test_collection(value, many=many)
-        return self._load(value, data, partial=partial, many=many)
+        return self._load(value, data, partial=partial, many=many, unknown=unknown)
 
 
 class Pluck(Nested):

src/marshmallow/schema.py

@@ -652,6 +652,7 @@ def _deserialize(
                     d_kwargs["partial"] = sub_partial
                 else:
                     d_kwargs["partial"] = partial
+                d_kwargs["unknown"] = unknown
                 getter = lambda val: field_obj.deserialize(
                     val, field_name, data, **d_kwargs
                 )
@@ -665,6 +666,7 @@ def _deserialize(
                 if value is not missing:
                     key = field_obj.attribute or attr_name
                     set_value(typing.cast(typing.Dict, ret), key, value)
+            unknown = unknown or self.unknown
             if unknown != EXCLUDE:
                 fields = {
                     field_obj.data_key if field_obj.data_key is not None else field_name
@@ -701,14 +703,17 @@ def load(
             will be ignored. Use dot delimiters to specify nested fields.
         :param unknown: Whether to exclude, include, or raise an error for unknown
             fields in the data. Use `EXCLUDE`, `INCLUDE` or `RAISE`.
-            If `None`, the value for `self.unknown` is used.
+            If `None`, the value for `self.unknown` is used. When used, provided
+            value is propagated to the loading of nested schemas.
         :return: Deserialized data
 
         .. versionadded:: 1.0.0
         .. versionchanged:: 3.0.0b7
             This method returns the deserialized data rather than a ``(data, errors)`` duple.
             A :exc:`ValidationError <marshmallow.exceptions.ValidationError>` is raised
             if invalid data are passed.
+        .. versionchanged:: 3.2.2
+            The ``unknown`` parameter value is propagated to the loading of nested schemas.
         """
         return self._do_load(
             data, many=many, partial=partial, unknown=unknown, postprocess=True
@@ -823,7 +828,6 @@ def _do_load(
         error_store = ErrorStore()
         errors = {}  # type: typing.Dict[str, typing.List[str]]
         many = self.many if many is None else bool(many)
-        unknown = unknown or self.unknown
         if partial is None:
             partial = self.partial
         # Run preprocessors

tests/test_schema.py

@@ -289,6 +289,23 @@ class Outer(Schema):
     assert Outer().load({"list1": val, "list2": val}) == {"list1": [], "list2": []}
 
 
+@pytest.mark.parametrize(
+    "val",
+    (
+        {"inner": {"name": "name"}, "unknown": 1},
+        {"inner": {"name": "name", "unknown_nested": 1}, "unknown": 1},
+    ),
+)
+def test_load_unknown(val):
+    class Inner(Schema):
+        name = fields.String()
+
+    class Outer(Schema):
+        inner = fields.Nested(Inner)
+
+    assert Outer().load(val, unknown=EXCLUDE) == {"inner": {"name": "name"}}
+
+
 def test_loads_returns_a_user():
     s = UserSchema()
     result = s.loads(json.dumps({"name": "Monty"}))