chore: add constructor fixture for tests (#373)

narwhals/_pandas_like/series.py

@@ -6,6 +6,7 @@
 from typing import Literal
 from typing import Sequence
 
+from narwhals._pandas_like.utils import int_dtype_mapper
 from narwhals._pandas_like.utils import native_series_from_iterable
 from narwhals._pandas_like.utils import reverse_translate_dtype
 from narwhals._pandas_like.utils import to_datetime
@@ -609,14 +610,23 @@ def second(self) -> PandasSeries:
         )
 
     def millisecond(self) -> PandasSeries:
+        if "pyarrow" in str(self._series._series.dtype):
+            msg = ".dt.millisecond not implemented for pyarrow-backed pandas"
+            raise NotImplementedError(msg)
         return self._series._from_series(
             self._series._series.dt.microsecond // 1000,
         )
 
     def microsecond(self) -> PandasSeries:
+        if "pyarrow" in str(self._series._series.dtype):
+            msg = ".dt.microsecond not implemented for pyarrow-backed pandas"
+            raise NotImplementedError(msg)
         return self._series._from_series(self._series._series.dt.microsecond)
 
     def nanosecond(self) -> PandasSeries:
+        if "pyarrow" in str(self._series._series.dtype):
+            msg = ".dt.nanosecond not implemented for pyarrow-backed pandas"
+            raise NotImplementedError(msg)
         return self._series._from_series(
             (
                 (self._series._series.dt.microsecond * 1_000)
@@ -639,54 +649,59 @@ def ordinal_day(self) -> PandasSeries:
     def total_minutes(self) -> PandasSeries:
         s = self._series._series.dt.total_seconds()
         s_sign = (
-            2 * (s > 0).astype(int) - 1
+            2 * (s > 0).astype(int_dtype_mapper(s.dtype)) - 1
         )  # this calculates the sign of each series element
         s_abs = s.abs() // 60
         if ~s.isna().any():
-            s_abs = s_abs.astype(int)
+            s_abs = s_abs.astype(int_dtype_mapper(s.dtype))
         return self._series._from_series(s_abs * s_sign)
 
     def total_seconds(self) -> PandasSeries:
         s = self._series._series.dt.total_seconds()
         s_sign = (
-            2 * (s > 0).astype(int) - 1
+            2 * (s > 0).astype(int_dtype_mapper(s.dtype)) - 1
         )  # this calculates the sign of each series element
         s_abs = s.abs() // 1
         if ~s.isna().any():
-            s_abs = s_abs.astype(int)
+            s_abs = s_abs.astype(int_dtype_mapper(s.dtype))
         return self._series._from_series(s_abs * s_sign)
 
     def total_milliseconds(self) -> PandasSeries:
         s = self._series._series.dt.total_seconds() * 1e3
         s_sign = (
-            2 * (s > 0).astype(int) - 1
+            2 * (s > 0).astype(int_dtype_mapper(s.dtype)) - 1
         )  # this calculates the sign of each series element
         s_abs = s.abs() // 1
         if ~s.isna().any():
-            s_abs = s_abs.astype(int)
+            s_abs = s_abs.astype(int_dtype_mapper(s.dtype))
         return self._series._from_series(s_abs * s_sign)
 
     def total_microseconds(self) -> PandasSeries:
         s = self._series._series.dt.total_seconds() * 1e6
         s_sign = (
-            2 * (s > 0).astype(int) - 1
+            2 * (s > 0).astype(int_dtype_mapper(s.dtype)) - 1
         )  # this calculates the sign of each series element
         s_abs = s.abs() // 1
         if ~s.isna().any():
-            s_abs = s_abs.astype(int)
+            s_abs = s_abs.astype(int_dtype_mapper(s.dtype))
         return self._series._from_series(s_abs * s_sign)
 
     def total_nanoseconds(self) -> PandasSeries:
         s = self._series._series.dt.total_seconds() * 1e9
         s_sign = (
-            2 * (s > 0).astype(int) - 1
+            2 * (s > 0).astype(int_dtype_mapper(s.dtype)) - 1
         )  # this calculates the sign of each series element
         s_abs = s.abs() // 1
         if ~s.isna().any():
-            s_abs = s_abs.astype(int)
+            s_abs = s_abs.astype(int_dtype_mapper(s.dtype))
         return self._series._from_series(s_abs * s_sign)
 
     def to_string(self, format: str) -> PandasSeries:  # noqa: A002
         # Polars' parser treats `'%.f'` as pandas does `'.%f'`
-        format = format.replace("%.f", ".%f")
+        # PyArrow interprets `'%S'` as "seconds, plus fractional seconds"
+        # and doesn't support `%f`
+        if "pyarrow" not in str(self._series._series.dtype):
+            format = format.replace("%S%.f", "%S.%f")
+        else:
+            format = format.replace("%S.%f", "%S").replace("%S%.f", "%S")
         return self._series._from_series(self._series._series.dt.strftime(format))

narwhals/_pandas_like/utils.py

@@ -600,3 +600,11 @@ def to_datetime(implementation: str) -> Any:
     if implementation == "cudf":
         return get_cudf().to_datetime
     raise AssertionError
+
+
+def int_dtype_mapper(dtype: Any) -> str:
+    if "pyarrow" in str(dtype):
+        return "Int64[pyarrow]"
+    if str(dtype).lower() != str(dtype):  # pragma: no cover
+        return "Int64"
+    return "int64"

narwhals/dataframe.py

@@ -501,7 +501,7 @@ def drop_nulls(self) -> Self:
             >>> import polars as pl
             >>> import pandas as pd
             >>> import narwhals as nw
-            >>> data = {"a": [1.0, 2.0, None], "ba": [1, None, 2.0]}
+            >>> data = {"a": [1.0, 2.0, None], "ba": [1.0, None, 2.0]}
             >>> df_pd = pd.DataFrame(data)
             >>> df_pl = pl.DataFrame(data)
 
@@ -1828,9 +1828,7 @@ def pipe(self, function: Callable[[Any], Self], *args: Any, **kwargs: Any) -> Se
 
             >>> def func(df_any):
             ...     df = nw.from_native(df_any)
-            ...     df = df.pipe(
-            ...         lambda _df: _df.select([x for x in _df.columns if len(x) == 1])
-            ...     )
+            ...     df = df.pipe(lambda _df: _df.select("a"))
             ...     return nw.to_native(df)
 
             We can then pass either pandas or Polars:
@@ -1866,7 +1864,7 @@ def drop_nulls(self) -> Self:
             >>> import polars as pl
             >>> import pandas as pd
             >>> import narwhals as nw
-            >>> data = {"a": [1.0, 2.0, None], "ba": [1, None, 2.0]}
+            >>> data = {"a": [1.0, 2.0, None], "ba": [1.0, None, 2.0]}
             >>> df_pd = pd.DataFrame(data)
             >>> df_pl = pl.LazyFrame(data)
 
@@ -1972,7 +1970,7 @@ def columns(self) -> list[str]:
             ...     }
             ... ).select("foo", "bar")
             >>> lf = nw.from_native(lf_pl)
-            >>> lf.columns
+            >>> lf.columns  # doctest: +SKIP
             ['foo', 'bar']
         """
         return super().columns

narwhals/expression.py

@@ -2787,7 +2787,7 @@ def to_string(self, format: str) -> Expr:  # noqa: A002
 
             Therefore, we make the following adjustments:
 
-            - for pandas-like libraries, we replace `".%f"` with `"%.f"`.
+            - for pandas-like libraries, we replace `"%S.%f"` with `"%S%.f"`.
             - for PyArrow, we replace `"%S.%f"` with `"%S"`.
 
             Workarounds like these don't make us happy, and we try to avoid them as

narwhals/series.py

@@ -2775,7 +2775,7 @@ def to_string(self, format: str) -> Series:  # noqa: A002
 
             Therefore, we make the following adjustments:
 
-            - for pandas-like libraries, we replace `".%f"` with `"%.f"`.
+            - for pandas-like libraries, we replace `"%S.%f"` with `"%S%.f"`.
             - for PyArrow, we replace `"%S.%f"` with `"%S"`.
 
             Workarounds like these don't make us happy, and we try to avoid them as

tests/conftest.py

@@ -1,7 +1,16 @@
+import os
 from typing import Any
+from typing import Callable
 
+import pandas as pd
+import polars as pl
+import pyarrow as pa
 import pytest
 
+from narwhals.dependencies import get_modin
+from narwhals.typing import IntoDataFrame
+from narwhals.utils import parse_version
+
 
 def pytest_addoption(parser: Any) -> None:
     parser.addoption(
@@ -21,3 +30,44 @@ def pytest_collection_modifyitems(config: Any, items: Any) -> Any:  # pragma: no
     for item in items:
         if "slow" in item.keywords:
             item.add_marker(skip_slow)
+
+
+def pandas_constructor(obj: Any) -> IntoDataFrame:
+    return pd.DataFrame(obj)  # type: ignore[no-any-return]
+
+
+def pandas_nullable_constructor(obj: Any) -> IntoDataFrame:
+    return pd.DataFrame(obj).convert_dtypes()  # type: ignore[no-any-return]
+
+
+def pandas_pyarrow_constructor(obj: Any) -> IntoDataFrame:
+    return pd.DataFrame(obj).convert_dtypes(dtype_backend="pyarrow")  # type: ignore[no-any-return]
+
+
+def modin_constructor(obj: Any) -> IntoDataFrame:  # pragma: no cover
+    return pd.DataFrame(obj).convert_dtypes(dtype_backend="pyarrow")  # type: ignore[no-any-return]
+
+
+def polars_constructor(obj: Any) -> IntoDataFrame:
+    return pl.DataFrame(obj)
+
+
+if parse_version(pd.__version__) >= parse_version("1.5.0"):
+    params = [pandas_constructor, pandas_nullable_constructor, pandas_pyarrow_constructor]
+else:  # pragma: no cover
+    params = [pandas_constructor]
+params.append(polars_constructor)
+if os.environ.get("CI") and get_modin() is not None:  # pragma: no cover
+    params.append(modin_constructor)
+
+
+@pytest.fixture(params=params)
+def constructor(request: Any) -> Callable[[Any], IntoDataFrame]:
+    return request.param  # type: ignore[no-any-return]
+
+
+# TODO: once pyarrow has complete coverage, we can remove this one,
+# and just put `pa.table` into `constructor`
+@pytest.fixture(params=[*params, pa.table])
+def constructor_with_pyarrow(request: Any) -> Callable[[Any], IntoDataFrame]:
+    return request.param  # type: ignore[no-any-return]

tests/expr/any_all_test.py

@@ -0,0 +1,22 @@
+from typing import Any
+
+import narwhals as nw
+from tests.utils import compare_dicts
+
+
+def test_any_all(constructor: Any) -> None:
+    df = nw.from_native(
+        constructor(
+            {
+                "a": [True, False, True],
+                "b": [True, True, True],
+                "c": [False, False, False],
+            }
+        )
+    )
+    result = nw.to_native(df.select(nw.all().all()))
+    expected = {"a": [False], "b": [True], "c": [False]}
+    compare_dicts(result, expected)
+    result = nw.to_native(df.select(nw.all().any()))
+    expected = {"a": [True], "b": [True], "c": [False]}
+    compare_dicts(result, expected)

tests/expr/cat/get_categories_test.py

@@ -2,17 +2,12 @@
 
 from typing import Any
 
-import pandas as pd
-import polars as pl
-import pytest
-
 import narwhals as nw
 from tests.utils import compare_dicts
 
 data = {"a": ["one", "two", "two"]}
 
 
-@pytest.mark.parametrize("constructor", [pd.DataFrame, pl.DataFrame])
 def test_get_categories(constructor: Any) -> None:
     df = nw.from_native(constructor(data), eager_only=True)
     df = df.select(nw.col("a").cast(nw.Categorical))

tests/expr/cum_sum_test.py

@@ -1,10 +1,5 @@
 from typing import Any
 
-import pandas as pd
-import polars as pl
-import pyarrow as pa
-import pytest
-
 import narwhals as nw
 from tests.utils import compare_dicts
 
@@ -15,9 +10,8 @@
 }
 
 
-@pytest.mark.parametrize("constructor", [pd.DataFrame, pl.DataFrame, pa.table])
-def test_cum_sum_simple(constructor: Any) -> None:
-    df = nw.from_native(constructor(data), eager_only=True)
+def test_cum_sum_simple(constructor_with_pyarrow: Any) -> None:
+    df = nw.from_native(constructor_with_pyarrow(data), eager_only=True)
     result = df.select(nw.all().cum_sum())
     expected = {
         "a": [0, 1, 3, 6, 10],

tests/expr/diff_test.py

@@ -1,9 +1,5 @@
 from typing import Any
 
-import pandas as pd
-import polars as pl
-import pytest
-
 import narwhals as nw
 from tests.utils import compare_dicts
 
@@ -14,7 +10,6 @@
 }
 
 
-@pytest.mark.parametrize("constructor", [pd.DataFrame, pl.DataFrame])
 def test_over_single(constructor: Any) -> None:
     df = nw.from_native(constructor(data), eager_only=True)
     result = df.with_columns(c_diff=nw.col("c").diff()).filter(nw.col("i") > 0)

tests/expr/fill_null_test.py

@@ -1,9 +1,5 @@
 from typing import Any
 
-import pandas as pd
-import polars as pl
-import pytest
-
 import narwhals as nw
 from tests.utils import compare_dicts
 
@@ -14,7 +10,6 @@
 }
 
 
-@pytest.mark.parametrize("constructor", [pd.DataFrame, pl.DataFrame])
 def test_over_single(constructor: Any) -> None:
     df = nw.from_native(constructor(data), eager_only=True)
     result = df.with_columns(nw.all().fill_null(99))

tests/expr/filter_test.py

@@ -1,9 +1,5 @@
 from typing import Any
 
-import pandas as pd
-import polars as pl
-import pytest
-
 import narwhals as nw
 from tests.utils import compare_dicts
 
@@ -15,7 +11,6 @@
 }
 
 
-@pytest.mark.parametrize("constructor", [pd.DataFrame, pl.DataFrame])
 def test_filter(constructor: Any) -> None:
     df = nw.from_native(constructor(data), eager_only=True)
     result = df.select(nw.col("a").filter(nw.col("i") < 2, nw.col("c") == 5))

tests/expr/is_between_test.py

@@ -2,8 +2,6 @@
 
 from typing import Any
 
-import pandas as pd
-import polars as pl
 import pytest
 
 import narwhals as nw
@@ -14,7 +12,6 @@
 }
 
 
-@pytest.mark.parametrize("constructor", [pd.DataFrame, pl.DataFrame])
 @pytest.mark.parametrize(
     ("closed", "expected"),
     [

tests/expr/is_duplicated_test.py

@@ -1,9 +1,5 @@
 from typing import Any
 
-import pandas as pd
-import polars as pl
-import pytest
-
 import narwhals as nw
 from tests.utils import compare_dicts
 
@@ -13,7 +9,6 @@
 }
 
 
-@pytest.mark.parametrize("constructor", [pd.DataFrame, pl.DataFrame])
 def test_is_duplicated(constructor: Any) -> None:
     df = nw.from_native(constructor(data), eager_only=True)
     result = df.select(nw.all().is_duplicated())