fix(python): address DataFrame construction error with lists of numpy arrays

py-polars/polars/io/spreadsheet/functions.py

@@ -445,6 +445,10 @@ def _read_spreadsheet(
         if hasattr(parser, "close"):
             parser.close()
 
+    if not parsed_sheets:
+        param, value = ("id", sheet_id) if sheet_name is None else ("name", sheet_name)
+        raise ValueError(f"no matching sheets found when `sheet_{param}` is {value!r}")
+
     if return_multi:
         return parsed_sheets
     return next(iter(parsed_sheets.values()))

py-polars/polars/utils/_construction.py

@@ -1024,7 +1024,11 @@ def _sequence_of_sequence_to_pydf(
         local_schema_override = (
             include_unknowns(schema_overrides, column_names) if schema_overrides else {}
         )
-        if column_names and first_element and len(first_element) != len(column_names):
+        if (
+            column_names
+            and len(first_element) > 0
+            and len(first_element) != len(column_names)
+        ):
             raise ShapeError("the row data does not match the number of columns")
 
         unpack_nested = False

py-polars/tests/unit/io/test_spreadsheet.py

@@ -194,6 +194,36 @@ def test_read_excel_basic_datatypes(
         assert_frame_equal(df, df)
 
 
+@pytest.mark.parametrize(
+    ("read_spreadsheet", "source", "params"),
+    [
+        (pl.read_excel, "path_xlsx", {"engine": "xlsx2csv"}),
+        (pl.read_excel, "path_xlsx", {"engine": "openpyxl"}),
+        (pl.read_excel, "path_xlsb", {"engine": "pyxlsb"}),
+        (pl.read_ods, "path_ods", {}),
+    ],
+)
+def test_read_invalid_worksheet(
+    read_spreadsheet: Callable[..., dict[str, pl.DataFrame]],
+    source: str,
+    params: dict[str, str],
+    request: pytest.FixtureRequest,
+) -> None:
+    spreadsheet_path = request.getfixturevalue(source)
+    for param, sheet_id, sheet_name in (
+        ("id", 999, None),
+        ("name", None, "not_a_sheet_name"),
+    ):
+        value = sheet_id if param == "id" else sheet_name
+        with pytest.raises(
+            ValueError,
+            match=f"no matching sheets found when `sheet_{param}` is {value!r}",
+        ):
+            read_spreadsheet(
+                spreadsheet_path, sheet_id=sheet_id, sheet_name=sheet_name, **params
+            )
+
+
 @pytest.mark.parametrize("engine", ["xlsx2csv", "openpyxl"])
 def test_write_excel_bytes(engine: Literal["xlsx2csv", "openpyxl", "pyxlsb"]) -> None:
     df = pl.DataFrame({"A": [1, 2, 3, 4, 5]})

py-polars/tests/unit/test_constructors.py

@@ -592,6 +592,10 @@ def test_init_ndarray(monkeypatch: Any) -> None:
     assert df.shape == (2, 1)
     assert df.rows() == [([0, 1, 2, 3, 4],), ([5, 6, 7, 8, 9],)]
 
+    test_rows = [(1, 2), (3, 4)]
+    df = pl.DataFrame([np.array(test_rows[0]), np.array(test_rows[1])], orient="row")
+    assert_frame_equal(df, pl.DataFrame(test_rows, orient="row"))
+
     # numpy arrays containing NaN
     df0 = pl.DataFrame(
         data={"x": [1.0, 2.5, float("nan")], "y": [4.0, float("nan"), 6.5]},

py-polars/tests/unit/test_errors.py

@@ -105,7 +105,7 @@ def test_string_numeric_comp_err() -> None:
 def test_panic_error() -> None:
     with pytest.raises(
         pl.PolarsPanicError,
-        match="""dimensions cannot be empty""",
+        match="dimensions cannot be empty",
     ):
         pl.Series("a", [1, 2, 3]).reshape(())