fix: consider cells with invalid values as nulls (#223)

closes #220

Signed-off-by: Luka Peschke <[email protected]>

python/tests/test_fastexcel.py

@@ -548,3 +548,21 @@ def test_null_strings(excel_file: str):
         pl.col("TIMESTAMPS_AND_NULLS").dt.cast_time_unit("ms"),
     )
     pl_assert_frame_equal(sheet.to_polars(), pl_df)
+
+
+def test_null_values_in_cells() -> None:
+    excel_reader = fastexcel.read_excel(path_for_fixture("fixture-invalid-cell-value.xlsx"))
+    sheet = excel_reader.load_sheet(0)
+
+    expected = {
+        "Title": ["A", "B", "C", "D"],
+        "Date": [None, None, datetime(2021, 1, 1), datetime(2021, 5, 5)],
+    }
+
+    pl_assert_frame_equal(
+        sheet.to_polars(),
+        pl.DataFrame(expected).with_columns(pl.col("Date").dt.cast_time_unit("ms")),
+    )
+    pd_expected = pd.DataFrame(expected)
+    pd_expected["Date"] = pd_expected["Date"].dt.as_unit("ms")
+    pd_assert_frame_equal(sheet.to_pandas(), pd_expected)

src/types/dtype.rs

@@ -136,7 +136,7 @@ fn get_cell_dtype(data: &Range<CalData>, row: usize, col: usize) -> FastExcelRes
         CalData::DurationIso(_) => Ok(DType::Duration),
         // Errors and nulls
         CalData::Error(err) => match err {
-            CellErrorType::NA => Ok(DType::Null),
+            CellErrorType::NA | CellErrorType::Value | CellErrorType::Null => Ok(DType::Null),
             _ => Err(FastExcelErrorKind::CalamineCellError(err.to_owned()).into()),
         },
         CalData::Empty => Ok(DType::Null),