feat[python]: Complain when Pandas Dataframe contains duplicated colu…

…mn names (#4733) (#4744)

py-polars/polars/internals/construction.py

@@ -354,7 +354,7 @@ def _pandas_series_to_arrow(
     Arrow Array
 
     """
-    dtype = values.dtype
+    dtype = getattr(values, "dtype", None)
     if dtype == "object" and len(values) > 0:
         first_non_none = _get_first_non_none(values.values)  # type: ignore[arg-type]
 
@@ -364,8 +364,15 @@ def _pandas_series_to_arrow(
             return pa.nulls(min_len, pa.large_utf8())
 
         return pa.array(values, from_pandas=nan_to_none)
-    else:
+    elif dtype:
         return pa.array(values, from_pandas=nan_to_none)
+    else:
+        # Pandas Series is actually a Pandas DataFrame when the original dataframe
+        # contains duplicated columns and a duplicated column is requested with df["a"].
+        raise ValueError(
+            "Duplicate column names found: "
+            + f"{str(values.columns.tolist())}"  # type: ignore[union-attr]
+        )
 
 
 def pandas_to_pyseries(

py-polars/tests/unit/test_interop.py

@@ -153,6 +153,12 @@ def test_from_pandas_datetime() -> None:
     pl.from_pandas(df)
 
 
+def test_from_pandas_duplicated_columns() -> None:
+    df = pd.DataFrame([[1, 2, 3, 4], [5, 6, 7, 8]], columns=["a", "b", "c", "b"])
+    with pytest.raises(ValueError, match="Duplicate column names found: "):
+        pl.from_pandas(df)
+
+
 def test_arrow_list_roundtrip() -> None:
     # https://github.com/pola-rs/polars/issues/1064
     tbl = pa.table({"a": [1], "b": [[1, 2]]})