Fixed #38419 - BUG: set_index screws up the dtypes on empty DataFrames

Author: jordi-crespo

pandas/tests/frame/methods/test_set_index.py

@@ -49,13 +49,14 @@ def test_set_index_empty_column(self):
 
     def test_set_index_empty_dataframe(self):
         # GH#38419
-        df = DataFrame(
+        df1 = DataFrame(
             {"a": Series(dtype="datetime64[ns]"), "b": Series(dtype="int64"), "c": []}
         )
-
-        expected = df.set_index(["a", "b"])
-        tm.assert_frame_equal(df.reset_index(drop=True), expected.reset_index())
-        assert df.empty is True
+    
+        df2 = df1.set_index(['a', 'b'])
+        result = df2.index.to_frame().dtypes
+        expected = df1[['a', 'b']].dtypes
+        tm.assert_series_equal(result, expected)
 
     def test_set_index_multiindexcolumns(self):
         columns = MultiIndex.from_tuples([("foo", 1), ("foo", 2), ("bar", 1)])
@@ -153,9 +154,9 @@ def test_set_index_cast(self):
         tm.assert_frame_equal(df, df2)
 
     # A has duplicate values, C does not
-    @pytest.mark.parametrize("keys", ["A", "C", ["A", "B"], ("tuple", "as", "label")])
-    @pytest.mark.parametrize("inplace", [True, False])
-    @pytest.mark.parametrize("drop", [True, False])
+    @ pytest.mark.parametrize("keys", ["A", "C", ["A", "B"], ("tuple", "as", "label")])
+    @ pytest.mark.parametrize("inplace", [True, False])
+    @ pytest.mark.parametrize("drop", [True, False])
     def test_set_index_drop_inplace(self, frame_of_index_cols, drop, inplace, keys):
         df = frame_of_index_cols
 
@@ -176,8 +177,8 @@ def test_set_index_drop_inplace(self, frame_of_index_cols, drop, inplace, keys):
         tm.assert_frame_equal(result, expected)
 
     # A has duplicate values, C does not
-    @pytest.mark.parametrize("keys", ["A", "C", ["A", "B"], ("tuple", "as", "label")])
-    @pytest.mark.parametrize("drop", [True, False])
+    @ pytest.mark.parametrize("keys", ["A", "C", ["A", "B"], ("tuple", "as", "label")])
+    @ pytest.mark.parametrize("drop", [True, False])
     def test_set_index_append(self, frame_of_index_cols, drop, keys):
         df = frame_of_index_cols
 
@@ -193,8 +194,8 @@ def test_set_index_append(self, frame_of_index_cols, drop, keys):
         tm.assert_frame_equal(result, expected)
 
     # A has duplicate values, C does not
-    @pytest.mark.parametrize("keys", ["A", "C", ["A", "B"], ("tuple", "as", "label")])
-    @pytest.mark.parametrize("drop", [True, False])
+    @ pytest.mark.parametrize("keys", ["A", "C", ["A", "B"], ("tuple", "as", "label")])
+    @ pytest.mark.parametrize("drop", [True, False])
     def test_set_index_append_to_multiindex(self, frame_of_index_cols, drop, keys):
         # append to existing multiindex
         df = frame_of_index_cols.set_index(["D"], drop=drop, append=True)
@@ -218,7 +219,7 @@ def test_set_index_after_mutation(self):
     # MultiIndex constructor does not work directly on Series -> lambda
     # Add list-of-list constructor because list is ambiguous -> lambda
     # also test index name if append=True (name is duplicate here for B)
-    @pytest.mark.parametrize(
+    @ pytest.mark.parametrize(
         "box",
         [
             Series,
@@ -229,10 +230,10 @@ def test_set_index_after_mutation(self):
             lambda x: MultiIndex.from_arrays([x]),
         ],
     )
-    @pytest.mark.parametrize(
+    @ pytest.mark.parametrize(
         "append, index_name", [(True, None), (True, "B"), (True, "test"), (False, None)]
     )
-    @pytest.mark.parametrize("drop", [True, False])
+    @ pytest.mark.parametrize("drop", [True, False])
     def test_set_index_pass_single_array(
         self, frame_of_index_cols, drop, append, index_name, box
     ):
@@ -261,14 +262,14 @@ def test_set_index_pass_single_array(
 
     # MultiIndex constructor does not work directly on Series -> lambda
     # also test index name if append=True (name is duplicate here for A & B)
-    @pytest.mark.parametrize(
+    @ pytest.mark.parametrize(
         "box", [Series, Index, np.array, list, lambda x: MultiIndex.from_arrays([x])]
     )
-    @pytest.mark.parametrize(
+    @ pytest.mark.parametrize(
         "append, index_name",
         [(True, None), (True, "A"), (True, "B"), (True, "test"), (False, None)],
     )
-    @pytest.mark.parametrize("drop", [True, False])
+    @ pytest.mark.parametrize("drop", [True, False])
     def test_set_index_pass_arrays(
         self, frame_of_index_cols, drop, append, index_name, box
     ):
@@ -292,7 +293,7 @@ def test_set_index_pass_arrays(
     # MultiIndex constructor does not work directly on Series -> lambda
     # We also emulate a "constructor" for the label -> lambda
     # also test index name if append=True (name is duplicate here for A)
-    @pytest.mark.parametrize(
+    @ pytest.mark.parametrize(
         "box2",
         [
             Series,
@@ -304,7 +305,7 @@ def test_set_index_pass_arrays(
             lambda x: x.name,
         ],
     )
-    @pytest.mark.parametrize(
+    @ pytest.mark.parametrize(
         "box1",
         [
             Series,
@@ -316,10 +317,10 @@ def test_set_index_pass_arrays(
             lambda x: x.name,
         ],
     )
-    @pytest.mark.parametrize(
+    @ pytest.mark.parametrize(
         "append, index_name", [(True, None), (True, "A"), (True, "test"), (False, None)]
     )
-    @pytest.mark.parametrize("drop", [True, False])
+    @ pytest.mark.parametrize("drop", [True, False])
     def test_set_index_pass_arrays_duplicate(
         self, frame_of_index_cols, drop, append, index_name, box1, box2
     ):
@@ -352,8 +353,8 @@ def test_set_index_pass_arrays_duplicate(
         expected = expected.set_index([keys[1]], drop=drop, append=True)
         tm.assert_frame_equal(result, expected)
 
-    @pytest.mark.parametrize("append", [True, False])
-    @pytest.mark.parametrize("drop", [True, False])
+    @ pytest.mark.parametrize("append", [True, False])
+    @ pytest.mark.parametrize("drop", [True, False])
     def test_set_index_pass_multiindex(self, frame_of_index_cols, drop, append):
         df = frame_of_index_cols
         keys = MultiIndex.from_arrays([df["A"], df["B"]], names=["A", "B"])
@@ -515,8 +516,8 @@ def test_set_index_verify_integrity(self, frame_of_index_cols):
         with pytest.raises(ValueError, match="Index has duplicate keys"):
             df.set_index([df["A"], df["A"]], verify_integrity=True)
 
-    @pytest.mark.parametrize("append", [True, False])
-    @pytest.mark.parametrize("drop", [True, False])
+    @ pytest.mark.parametrize("append", [True, False])
+    @ pytest.mark.parametrize("drop", [True, False])
     def test_set_index_raise_keys(self, frame_of_index_cols, drop, append):
         df = frame_of_index_cols
 
@@ -537,9 +538,9 @@ def test_set_index_raise_keys(self, frame_of_index_cols, drop, append):
         with pytest.raises(KeyError, match=msg):
             df.set_index(["A", df["A"], tuple(df["A"])], drop=drop, append=append)
 
-    @pytest.mark.parametrize("append", [True, False])
-    @pytest.mark.parametrize("drop", [True, False])
-    @pytest.mark.parametrize("box", [set], ids=["set"])
+    @ pytest.mark.parametrize("append", [True, False])
+    @ pytest.mark.parametrize("drop", [True, False])
+    @ pytest.mark.parametrize("box", [set], ids=["set"])
     def test_set_index_raise_on_type(self, frame_of_index_cols, box, drop, append):
         df = frame_of_index_cols
 
@@ -553,14 +554,14 @@ def test_set_index_raise_on_type(self, frame_of_index_cols, box, drop, append):
             df.set_index(["A", df["A"], box(df["A"])], drop=drop, append=append)
 
     # MultiIndex constructor does not work directly on Series -> lambda
-    @pytest.mark.parametrize(
+    @ pytest.mark.parametrize(
         "box",
         [Series, Index, np.array, iter, lambda x: MultiIndex.from_arrays([x])],
         ids=["Series", "Index", "np.array", "iter", "MultiIndex"],
     )
-    @pytest.mark.parametrize("length", [4, 6], ids=["too_short", "too_long"])
-    @pytest.mark.parametrize("append", [True, False])
-    @pytest.mark.parametrize("drop", [True, False])
+    @ pytest.mark.parametrize("length", [4, 6], ids=["too_short", "too_long"])
+    @ pytest.mark.parametrize("append", [True, False])
+    @ pytest.mark.parametrize("drop", [True, False])
     def test_set_index_raise_on_len(
         self, frame_of_index_cols, box, length, drop, append
     ):