fixes

Author: jorisvandenbossche

pandas/core/internals/blocks.py

@@ -929,7 +929,7 @@ def replace(
                     blocks = blk.convert(
                         copy=False,
                         using_cow=using_cow,
-                        convert_string=convert_string or self.dtype != _dtype_obj,
+                        convert_string=convert_string or self.dtype == "string",
                     )
                     if len(blocks) > 1 or blocks[0].dtype != blk.dtype:
                         warnings.warn(
@@ -987,7 +987,7 @@ def _replace_regex(
         inplace: bool = False,
         mask=None,
         using_cow: bool = False,
-        convert_string: bool = True,
+        convert_string=None,
         already_warned=None,
     ) -> list[Block]:
         """
@@ -1048,10 +1048,18 @@ def _replace_regex(
                 already_warned.warned_already = True
 
         nbs = block.convert(
-            copy=False, using_cow=using_cow, convert_string=convert_string
+            copy=False,
+            using_cow=using_cow,
+            convert_string=convert_string or self.dtype == "string",
         )
         opt = get_option("future.no_silent_downcasting")
-        if (len(nbs) > 1 or nbs[0].dtype != block.dtype) and not opt:
+        if (
+            len(nbs) > 1
+            or (
+                nbs[0].dtype != block.dtype
+                and not (self.dtype == "string" and nbs[0].dtype == "string")
+            )
+        ) and not opt:
             warnings.warn(
                 # GH#54710
                 "Downcasting behavior in `replace` is deprecated and "
@@ -1088,7 +1096,7 @@ def replace_list(
             values._replace(to_replace=src_list, value=dest_list, inplace=True)
             return [blk]
 
-        convert_string = self.dtype != _dtype_obj
+        convert_string = self.dtype == "string"
 
         # Exclude anything that we know we won't contain
         pairs = [

pandas/tests/frame/methods/test_replace.py

@@ -299,7 +299,7 @@ def test_regex_replace_dict_nested_gh4115(self):
 
         tm.assert_frame_equal(result, expected)
 
-    def test_regex_replace_list_to_scalar(self, mix_abc):
+    def test_regex_replace_list_to_scalar(self, mix_abc, using_infer_string):
         df = DataFrame(mix_abc)
         expec = DataFrame(
             {
@@ -308,17 +308,20 @@ def test_regex_replace_list_to_scalar(self, mix_abc):
                 "c": [np.nan, np.nan, np.nan, "d"],
             }
         )
+        if using_infer_string:
+            expec["b"] = expec["b"].astype("str")
         msg = "Downcasting behavior in `replace`"
-        with tm.assert_produces_warning(FutureWarning, match=msg):
+        warn = None if using_infer_string else FutureWarning
+        with tm.assert_produces_warning(warn, match=msg):
             res = df.replace([r"\s*\.\s*", "a|b"], np.nan, regex=True)
         res2 = df.copy()
         res3 = df.copy()
-        with tm.assert_produces_warning(FutureWarning, match=msg):
+        with tm.assert_produces_warning(warn, match=msg):
             return_value = res2.replace(
                 [r"\s*\.\s*", "a|b"], np.nan, regex=True, inplace=True
             )
         assert return_value is None
-        with tm.assert_produces_warning(FutureWarning, match=msg):
+        with tm.assert_produces_warning(warn, match=msg):
             return_value = res3.replace(
                 regex=[r"\s*\.\s*", "a|b"], value=np.nan, inplace=True
             )
@@ -338,8 +341,6 @@ def test_regex_replace_str_to_numeric(self, mix_abc):
         return_value = res3.replace(regex=r"\s*\.\s*", value=0, inplace=True)
         assert return_value is None
         expec = DataFrame({"a": mix_abc["a"], "b": ["a", "b", 0, 0], "c": mix_abc["c"]})
-        # TODO(infer_string)
-        expec["c"] = expec["c"].astype(object)
         tm.assert_frame_equal(res, expec)
         tm.assert_frame_equal(res2, expec)
         tm.assert_frame_equal(res3, expec)
@@ -626,11 +627,7 @@ def test_replace_mixed2(self, using_infer_string):
                 "B": Series([0, "foo"], dtype="object"),
             }
         )
-        if using_infer_string:
-            with tm.assert_produces_warning(FutureWarning, match="Downcasting"):
-                result = df.replace([1, 2], ["foo", "bar"])
-        else:
-            result = df.replace([1, 2], ["foo", "bar"])
+        result = df.replace([1, 2], ["foo", "bar"])
         tm.assert_frame_equal(result, expected)
 
     def test_replace_mixed3(self):
@@ -1513,13 +1510,11 @@ def test_replace_with_compiled_regex(self):
         expected = DataFrame(["z", "b", "c"])
         tm.assert_frame_equal(result, expected)
 
-    def test_replace_intervals(self, using_infer_string):
+    def test_replace_intervals(self):
         # https://github.com/pandas-dev/pandas/issues/35931
         df = DataFrame({"a": [pd.Interval(0, 1), pd.Interval(0, 1)]})
-        warning = FutureWarning if using_infer_string else None
-        with tm.assert_produces_warning(warning, match="Downcasting"):
-            result = df.replace({"a": {pd.Interval(0, 1): "x"}})
-        expected = DataFrame({"a": ["x", "x"]})
+        result = df.replace({"a": {pd.Interval(0, 1): "x"}})
+        expected = DataFrame({"a": ["x", "x"]}, dtype=object)
         tm.assert_frame_equal(result, expected)
 
     def test_replace_unicode(self):
@@ -1620,7 +1615,7 @@ def test_regex_replace_scalar(
         tm.assert_frame_equal(result, expected)
 
     @pytest.mark.parametrize("regex", [False, True])
-    def test_replace_regex_dtype_frame(self, regex):
+    def test_replace_regex_dtype_frame(self, regex, using_infer_string):
         # GH-48644
         df1 = DataFrame({"A": ["0"], "B": ["0"]})
         expected_df1 = DataFrame({"A": [1], "B": [1]})