Preserve index when setting new column on empty dataframe. (#26471)

Author: sumanau7

pandas/core/frame.py

@@ -3098,7 +3098,7 @@ def _ensure_valid_index(self, value):
         passed value.
         """
         # GH5632, make sure that we are a Series convertible
-        if not len(self.index) and is_list_like(value):
+        if not len(self.index) and is_list_like(value) and len(value):
             try:
                 value = Series(value)
             except (ValueError, NotImplementedError, TypeError):

pandas/tests/frame/test_indexing.py

@@ -821,6 +821,14 @@ def test_setitem_empty_frame_with_boolean(self, dtype, kwargs):
         df[df > df2] = 47
         assert_frame_equal(df, df2)
 
+    def test_setitem_with_empty_listlike(self):
+        # GH #17101
+        index = pd.Index([], name="idx")
+        result = pd.DataFrame(columns=["A"], index=index)
+        result["A"] = []
+        expected = pd.DataFrame(columns=["A"], index=index)
+        tm.assert_index_equal(result.index, expected.index)
+
     def test_setitem_scalars_no_index(self):
         # GH16823 / 17894
         df = DataFrame()

pandas/tests/indexing/test_partial.py

@@ -442,10 +442,10 @@ def test_partial_set_empty_frame(self):
         # these work as they don't really change
         # anything but the index
         # GH5632
-        expected = DataFrame(columns=["foo"], index=Index([], dtype="int64"))
+        expected = DataFrame(columns=["foo"], index=Index([], dtype="object"))
 
         def f():
-            df = DataFrame()
+            df = DataFrame(index=Index([], dtype="object"))
             df["foo"] = Series([], dtype="object")
             return df
 
@@ -469,22 +469,21 @@ def f():
         expected["foo"] = expected["foo"].astype("float64")
 
         def f():
-            df = DataFrame()
+            df = DataFrame(index=Index([], dtype="int64"))
             df["foo"] = []
             return df
 
         tm.assert_frame_equal(f(), expected)
 
         def f():
-            df = DataFrame()
+            df = DataFrame(index=Index([], dtype="int64"))
             df["foo"] = Series(np.arange(len(df)), dtype="float64")
             return df
 
         tm.assert_frame_equal(f(), expected)
 
         def f():
-            df = DataFrame()
-            tm.assert_index_equal(df.index, Index([], dtype="object"))
+            df = DataFrame(index=Index([], dtype="int64"))
             df["foo"] = range(len(df))
             return df