Backport PR #48890 on branch 1.5.x (ERR: Improve error message when assigning a complete row using 'at' m…) (#49310)

Backport PR #48890: ERR: Improve error message when assigning a complete row using 'at' m…

Co-authored-by: Ambuj Pawar <[email protected]>

Author: meeseeksmachine

pandas/core/frame.py

@@ -86,6 +86,7 @@
     function as nv,
     np_percentile_argname,
 )
+from pandas.errors import InvalidIndexError
 from pandas.util._decorators import (
     Appender,
     Substitution,
@@ -4220,6 +4221,13 @@ def _set_value(
                 self.loc[index, col] = value
             self._item_cache.pop(col, None)
 
+        except InvalidIndexError as ii_err:
+            # GH48729: Seems like you are trying to assign a value to a
+            # row when only scalar options are permitted
+            raise InvalidIndexError(
+                f"You can only assign a scalar value not a {type(value)}"
+            ) from ii_err
+
     def _ensure_valid_index(self, value) -> None:
         """
         Ensure that if we don't have an index, that we can create one from the

pandas/tests/indexing/test_at.py

@@ -197,8 +197,12 @@ def test_at_frame_raises_key_error2(self, indexer_al):
     def test_at_frame_multiple_columns(self):
         # GH#48296 - at shouldn't modify multiple columns
         df = DataFrame({"a": [1, 2], "b": [3, 4]})
-        with pytest.raises(InvalidIndexError, match=r"slice\(None, None, None\)"):
-            df.at[5] = [6, 7]
+        new_row = [6, 7]
+        with pytest.raises(
+            InvalidIndexError,
+            match=f"You can only assign a scalar value not a \\{type(new_row)}",
+        ):
+            df.at[5] = new_row
 
     def test_at_getitem_mixed_index_no_fallback(self):
         # GH#19860
@@ -220,3 +224,13 @@ def test_at_categorical_integers(self):
             for key in [0, 1]:
                 with pytest.raises(KeyError, match=str(key)):
                     df.at[key, key]
+
+    def test_at_applied_for_rows(self):
+        # GH#48729 .at should raise InvalidIndexError when assigning rows
+        df = DataFrame(index=["a"], columns=["col1", "col2"])
+        new_row = [123, 15]
+        with pytest.raises(
+            InvalidIndexError,
+            match=f"You can only assign a scalar value not a \\{type(new_row)}",
+        ):
+            df.at["a"] = new_row