TST: make tests stricter (#32527)

Author: jbrockmendel

pandas/tests/arithmetic/test_numeric.py

@@ -913,13 +913,13 @@ def test_frame_operators(self, float_frame):
 
     # TODO: taken from tests.series.test_operators; needs cleanup
     def test_series_operators(self):
-        def _check_op(series, other, op, pos_only=False, check_dtype=True):
+        def _check_op(series, other, op, pos_only=False):
             left = np.abs(series) if pos_only else series
             right = np.abs(other) if pos_only else other
 
             cython_or_numpy = op(left, right)
             python = left.combine(right, op)
-            tm.assert_series_equal(cython_or_numpy, python, check_dtype=check_dtype)
+            tm.assert_series_equal(cython_or_numpy, python)
 
         def check(series, other):
             simple_ops = ["add", "sub", "mul", "truediv", "floordiv", "mod"]
@@ -942,15 +942,15 @@ def check(series, other):
         check(tser, tser[::2])
         check(tser, 5)
 
-        def check_comparators(series, other, check_dtype=True):
-            _check_op(series, other, operator.gt, check_dtype=check_dtype)
-            _check_op(series, other, operator.ge, check_dtype=check_dtype)
-            _check_op(series, other, operator.eq, check_dtype=check_dtype)
-            _check_op(series, other, operator.lt, check_dtype=check_dtype)
-            _check_op(series, other, operator.le, check_dtype=check_dtype)
+        def check_comparators(series, other):
+            _check_op(series, other, operator.gt)
+            _check_op(series, other, operator.ge)
+            _check_op(series, other, operator.eq)
+            _check_op(series, other, operator.lt)
+            _check_op(series, other, operator.le)
 
         check_comparators(tser, 5)
-        check_comparators(tser, tser + 1, check_dtype=False)
+        check_comparators(tser, tser + 1)
 
     # TODO: taken from tests.series.test_operators; needs cleanup
     def test_divmod(self):

pandas/tests/dtypes/cast/test_construct_from_scalar.py

@@ -15,6 +15,4 @@ def test_cast_1d_array_like_from_scalar_categorical():
     expected = Categorical(["a", "a"], categories=cats)
 
     result = construct_1d_arraylike_from_scalar("a", len(expected), cat_type)
-    tm.assert_categorical_equal(
-        result, expected, check_category_order=True, check_dtype=True
-    )
+    tm.assert_categorical_equal(result, expected)

pandas/tests/frame/test_constructors.py

@@ -916,10 +916,7 @@ def test_constructor_mrecarray(self):
         # from GH3479
 
         assert_fr_equal = functools.partial(
-            tm.assert_frame_equal,
-            check_index_type=True,
-            check_column_type=True,
-            check_frame_type=True,
+            tm.assert_frame_equal, check_index_type=True, check_column_type=True,
         )
         arrays = [
             ("float", np.array([1.5, 2.0])),

pandas/tests/io/excel/test_writers.py

@@ -452,7 +452,7 @@ def test_float_types(self, np_type, path):
         reader = ExcelFile(path)
         recons = pd.read_excel(reader, "test1", index_col=0).astype(np_type)
 
-        tm.assert_frame_equal(df, recons, check_dtype=False)
+        tm.assert_frame_equal(df, recons)
 
     @pytest.mark.parametrize("np_type", [np.bool8, np.bool_])
     def test_bool_types(self, np_type, path):

pandas/tests/io/json/test_pandas.py

@@ -572,7 +572,6 @@ def test_blocks_compat_GH9037(self):
             df_roundtrip,
             check_index_type=True,
             check_column_type=True,
-            check_frame_type=True,
             by_blocks=True,
             check_exact=True,
         )

pandas/tests/io/json/test_ujson.py

@@ -48,6 +48,18 @@ def numpy(request):
     return request.param
 
 
+def get_int32_compat_dtype(numpy, orient):
+    # See GH#32527
+    dtype = np.int64
+    if not ((numpy is None or orient == "index") or (numpy is True and orient is None)):
+        if compat.is_platform_windows():
+            dtype = np.int32
+        else:
+            dtype = np.intp
+
+    return dtype
+
+
 class TestUltraJSONTests:
     @pytest.mark.skipif(
         compat.is_platform_32bit(), reason="not compliant on 32-bit, xref #15865"
@@ -833,13 +845,20 @@ def test_dataframe(self, orient, numpy):
         if orient == "records" and numpy:
             pytest.skip("Not idiomatic pandas")
 
+        dtype = get_int32_compat_dtype(numpy, orient)
+
         df = DataFrame(
-            [[1, 2, 3], [4, 5, 6]], index=["a", "b"], columns=["x", "y", "z"]
+            [[1, 2, 3], [4, 5, 6]],
+            index=["a", "b"],
+            columns=["x", "y", "z"],
+            dtype=dtype,
         )
         encode_kwargs = {} if orient is None else dict(orient=orient)
         decode_kwargs = {} if numpy is None else dict(numpy=numpy)
+        assert (df.dtypes == dtype).all()
 
         output = ujson.decode(ujson.encode(df, **encode_kwargs), **decode_kwargs)
+        assert (df.dtypes == dtype).all()
 
         # Ensure proper DataFrame initialization.
         if orient == "split":
@@ -857,7 +876,8 @@ def test_dataframe(self, orient, numpy):
         elif orient == "index":
             df = df.transpose()
 
-        tm.assert_frame_equal(output, df, check_dtype=False)
+        assert (df.dtypes == dtype).all()
+        tm.assert_frame_equal(output, df)
 
     def test_dataframe_nested(self, orient):
         df = DataFrame(
@@ -897,14 +917,20 @@ def test_dataframe_numpy_labelled(self, orient):
         tm.assert_frame_equal(output, df)
 
     def test_series(self, orient, numpy):
+        dtype = get_int32_compat_dtype(numpy, orient)
         s = Series(
-            [10, 20, 30, 40, 50, 60], name="series", index=[6, 7, 8, 9, 10, 15]
+            [10, 20, 30, 40, 50, 60],
+            name="series",
+            index=[6, 7, 8, 9, 10, 15],
+            dtype=dtype,
         ).sort_values()
+        assert s.dtype == dtype
 
         encode_kwargs = {} if orient is None else dict(orient=orient)
         decode_kwargs = {} if numpy is None else dict(numpy=numpy)
 
         output = ujson.decode(ujson.encode(s, **encode_kwargs), **decode_kwargs)
+        assert s.dtype == dtype
 
         if orient == "split":
             dec = _clean_dict(output)
@@ -920,7 +946,8 @@ def test_series(self, orient, numpy):
             s.name = None
             s.index = [0, 1, 2, 3, 4, 5]
 
-        tm.assert_series_equal(output, s, check_dtype=False)
+        assert s.dtype == dtype
+        tm.assert_series_equal(output, s)
 
     def test_series_nested(self, orient):
         s = Series(

pandas/tests/io/pytables/test_store.py

@@ -2298,9 +2298,7 @@ def test_index_types(self, setup_path):
         with catch_warnings(record=True):
             values = np.random.randn(2)
 
-            func = lambda l, r: tm.assert_series_equal(
-                l, r, check_dtype=True, check_index_type=True
-            )
+            func = lambda l, r: tm.assert_series_equal(l, r, check_index_type=True)
 
         with catch_warnings(record=True):
             ser = Series(values, [0, "y"])

pandas/tests/io/test_clipboard.py

@@ -75,7 +75,11 @@ def df(request):
         )
     elif data_type == "mixed":
         return DataFrame(
-            {"a": np.arange(1.0, 6.0) + 0.01, "b": np.arange(1, 6), "c": list("abcde")}
+            {
+                "a": np.arange(1.0, 6.0) + 0.01,
+                "b": np.arange(1, 6).astype(np.int64),
+                "c": list("abcde"),
+            }
         )
     elif data_type == "float":
         return tm.makeCustomDataframe(
@@ -146,7 +150,7 @@ class TestClipboard:
     def check_round_trip_frame(self, data, excel=None, sep=None, encoding=None):
         data.to_clipboard(excel=excel, sep=sep, encoding=encoding)
         result = read_clipboard(sep=sep or "\t", index_col=0, encoding=encoding)
-        tm.assert_frame_equal(data, result, check_dtype=False)
+        tm.assert_frame_equal(data, result)
 
     # Test that default arguments copy as tab delimited
     def test_round_trip_frame(self, df):

pandas/tests/reshape/merge/test_join.py

@@ -298,7 +298,7 @@ def test_join_on_inner(self):
 
         expected = df.join(df2, on="key")
         expected = expected[expected["value"].notna()]
-        tm.assert_series_equal(joined["key"], expected["key"], check_dtype=False)
+        tm.assert_series_equal(joined["key"], expected["key"])
         tm.assert_series_equal(joined["value"], expected["value"], check_dtype=False)
         tm.assert_index_equal(joined.index, expected.index)
 

pandas/tests/reshape/test_union_categoricals.py

@@ -41,7 +41,7 @@ def test_union_categorical(self):
             for box in [Categorical, CategoricalIndex, Series]:
                 result = union_categoricals([box(Categorical(a)), box(Categorical(b))])
                 expected = Categorical(combined)
-                tm.assert_categorical_equal(result, expected, check_category_order=True)
+                tm.assert_categorical_equal(result, expected)
 
         # new categories ordered by appearance
         s = Categorical(["x", "y", "z"])

pandas/tests/series/methods/test_argsort.py

@@ -49,8 +49,8 @@ def test_argsort_stable(self):
         mexpected = np.argsort(s.values, kind="mergesort")
         qexpected = np.argsort(s.values, kind="quicksort")
 
-        tm.assert_series_equal(mindexer, Series(mexpected), check_dtype=False)
-        tm.assert_series_equal(qindexer, Series(qexpected), check_dtype=False)
+        tm.assert_series_equal(mindexer.astype(np.intp), Series(mexpected))
+        tm.assert_series_equal(qindexer.astype(np.intp), Series(qexpected))
         msg = (
             r"ndarray Expected type <class 'numpy\.ndarray'>, "
             r"found <class 'pandas\.core\.series\.Series'> instead"

pandas/tests/series/methods/test_replace.py

@@ -294,7 +294,7 @@ def test_replace_categorical(self, categorical, numeric):
         s = pd.Series(categorical)
         result = s.replace({"A": 1, "B": 2})
         expected = pd.Series(numeric)
-        tm.assert_series_equal(expected, result, check_dtype=False)
+        tm.assert_series_equal(expected, result)
 
     def test_replace_categorical_single(self):
         # GH 26988

pandas/tests/series/test_duplicates.py

@@ -47,9 +47,9 @@ def test_unique():
 
     # GH 18051
     s = Series(Categorical([]))
-    tm.assert_categorical_equal(s.unique(), Categorical([]), check_dtype=False)
+    tm.assert_categorical_equal(s.unique(), Categorical([]))
     s = Series(Categorical([np.nan]))
-    tm.assert_categorical_equal(s.unique(), Categorical([np.nan]), check_dtype=False)
+    tm.assert_categorical_equal(s.unique(), Categorical([np.nan]))
 
 
 def test_unique_data_ownership():