From d46b3b94a76fae77d3825670fc2e0c5ffc67db2f Mon Sep 17 00:00:00 2001
From: Niels Bantilan <niels.bantilan@gmail.com>
Date: Tue, 26 Sep 2023 17:22:37 -0700
Subject: [PATCH] add unit tests (#1351)

* add unit tests

Signed-off-by: Niels Bantilan <niels.bantilan@gmail.com>

* update

Signed-off-by: Niels Bantilan <niels.bantilan@gmail.com>

---------

Signed-off-by: Niels Bantilan <niels.bantilan@gmail.com>
---
 pandera/engines/pandas_engine.py | 11 +++++--
 tests/core/test_dtypes.py        | 50 ++++++++++++++++++++++++--------
 2 files changed, 47 insertions(+), 14 deletions(-)

diff --git a/pandera/engines/pandas_engine.py b/pandera/engines/pandas_engine.py
index e55f617ce..f78c2c46f 100644
--- a/pandera/engines/pandas_engine.py
+++ b/pandera/engines/pandas_engine.py
@@ -30,6 +30,7 @@
 import pandas as pd
 import typeguard
 from pydantic import BaseModel, ValidationError, create_model
+from typeguard import CollectionCheckStrategy
 
 from pandera import dtypes, errors
 from pandera.dtypes import immutable
@@ -1157,7 +1158,7 @@ def _check_type(self, element: Any) -> bool:
         # if the element is None or pd.NA, this function should return True:
         # the schema should only fail if nullable=False is specifed at the
         # schema/schema component level.
-        if element is None or pd.isna(element):
+        if element is None or element is pd.NA:
             return True
 
         try:
@@ -1177,7 +1178,13 @@ def _check_type(self, element: Any) -> bool:
                 from typing import TypedDict as _TypedDict
 
                 _type = _TypedDict(_type.__name__, _type.__annotations__)  # type: ignore
-            typeguard.check_type(element, _type)
+
+            typeguard.check_type(
+                element,
+                _type,
+                # This may be worth making configurable at the global level.
+                collection_check_strategy=CollectionCheckStrategy.ALL_ITEMS,
+            )
             return True
         except typeguard.TypeCheckError:
             return False
diff --git a/tests/core/test_dtypes.py b/tests/core/test_dtypes.py
index bc3a9a25b..d38e76af9 100644
--- a/tests/core/test_dtypes.py
+++ b/tests/core/test_dtypes.py
@@ -790,12 +790,12 @@ def test_python_std_list_dict_generics():
     "data_dict",
     [
         {
-            "dict_column": [{"foo": 1, "bar": 2}, {}, None],
-            "list_column": [[1.0], [], None],
+            "dict_column": [{"foo": 1}, {"foo": 1, "bar": 2}, {}, None],
+            "list_column": [[1.0], [1.0, 2.0], [], None],
         },
         {
-            "dict_column": [{"foo": "1", "bar": "2"}, {}, None],
-            "list_column": [["1.0"], [], None],
+            "dict_column": [{"foo": "1"}, {"foo": "1", "bar": "2"}, {}, None],
+            "list_column": [["1.0"], ["1.0", "2.0"], [], None],
         },
     ],
 )
@@ -816,8 +816,8 @@ def test_python_typing_handle_empty_list_dict_and_none(nullable, data_dict):
 
     expected = pd.DataFrame(
         {
-            "dict_column": [{"foo": 1, "bar": 2}, {}, pd.NA],
-            "list_column": [[1.0], [], pd.NA],
+            "dict_column": [{"foo": 1}, {"foo": 1, "bar": 2}, {}, pd.NA],
+            "list_column": [[1.0], [1.0, 2.0], [], pd.NA],
         }
     )
 
@@ -838,12 +838,12 @@ def test_python_typing_handle_empty_list_dict_and_none(nullable, data_dict):
     "data_dict",
     [
         {
-            "dict_column": [{"foo": 1, "bar": 2}, {}, None],
-            "list_column": [[1.0], [], None],
+            "dict_column": [{"foo": 1}, {"foo": 1, "bar": 2}, {}, None],
+            "list_column": [[1.0], [1.0, 2.0], [], None],
         },
         {
-            "dict_column": [{"foo": "1", "bar": "2"}, {}, None],
-            "list_column": [["1.0"], [], None],
+            "dict_column": [{"foo": "1"}, {"foo": "1", "bar": "2"}, {}, None],
+            "list_column": [["1.0"], ["1.0", "2.0"], [], None],
         },
     ],
 )
@@ -865,8 +865,8 @@ def test_python_std_list_dict_empty_and_none(nullable, data_dict):
 
     expected = pd.DataFrame(
         {
-            "dict_column": [{"foo": 1, "bar": 2}, {}, pd.NA],
-            "list_column": [[1.0], [], pd.NA],
+            "dict_column": [{"foo": 1}, {"foo": 1, "bar": 2}, {}, pd.NA],
+            "list_column": [[1.0], [1.0, 2.0], [], pd.NA],
         }
     )
 
@@ -876,3 +876,29 @@ def test_python_std_list_dict_empty_and_none(nullable, data_dict):
     else:
         with pytest.raises(pa.errors.SchemaError):
             schema.validate(data)
+
+
+def test_python_std_list_dict_error():
+    """Test that non-standard dict/list invalid values raise Schema Error."""
+    schema = pa.DataFrameSchema(
+        {
+            "dict_column": pa.Column(Dict[str, int]),
+            "list_column": pa.Column(List[float]),
+        },
+    )
+
+    data = pd.DataFrame(
+        {
+            "dict_column": [{"foo": 1}, {"foo": 1, "bar": "2"}, {}],
+            "list_column": [[1.0], ["1.0", 2.0], []],
+        }
+    )
+
+    try:
+        schema.validate(data, lazy=True)
+    except pa.errors.SchemaErrors as exc:
+        assert exc.failure_cases["failure_case"].iloc[0] == {
+            "foo": 1,
+            "bar": "2",
+        }
+        assert exc.failure_cases["failure_case"].iloc[1] == ["1.0", 2.0]