dont run both modin and modin[pyarrow] in ci

Author: MarcoGorelli

narwhals/_arrow/expr.py

@@ -352,8 +352,8 @@ def is_first_distinct(self: Self) -> Self:
     def is_last_distinct(self: Self) -> Self:
         return reuse_series_implementation(self, "is_last_distinct")
 
-    def unique(self: Self, *, maintain_order: bool) -> Self:
-        return reuse_series_implementation(self, "unique", maintain_order=maintain_order)
+    def unique(self: Self) -> Self:
+        return reuse_series_implementation(self, "unique", maintain_order=False)
 
     def replace_strict(
         self: Self, old: Sequence[Any], new: Sequence[Any], *, return_dtype: DType | None

narwhals/_arrow/series.py

@@ -683,9 +683,7 @@ def is_sorted(self: Self, *, descending: bool) -> bool:
         return maybe_extract_py_scalar(result, return_py_scalar=True)  # type: ignore[no-any-return]
 
     def unique(self: Self, *, maintain_order: bool) -> ArrowSeries:
-        # The param `maintain_order` is only here for compatibility with the Polars API
-        # and has no effect on the output.
-
+        # TODO(marco): `pc.unique` seems to always maintain order, is that guaranteed?
         return self._from_native_series(pc.unique(self._native_series))
 
     def replace_strict(

narwhals/_dask/expr.py

@@ -447,8 +447,7 @@ def round(self, decimals: int) -> Self:
             returns_scalar=self._returns_scalar,
         )
 
-    def unique(self, *, maintain_order: bool) -> Self:
-        # TODO(marco): maintain_order has no effect and will be deprecated
+    def unique(self) -> Self:
         return self._from_call(
             lambda _input: _input.unique(),
             "unique",

narwhals/_pandas_like/expr.py

@@ -351,8 +351,8 @@ def abs(self) -> Self:
     def cum_sum(self: Self, *, reverse: bool) -> Self:
         return reuse_series_implementation(self, "cum_sum", reverse=reverse)
 
-    def unique(self, *, maintain_order: bool = False) -> Self:
-        return reuse_series_implementation(self, "unique", maintain_order=maintain_order)
+    def unique(self) -> Self:
+        return reuse_series_implementation(self, "unique", maintain_order=False)
 
     def diff(self) -> Self:
         return reuse_series_implementation(self, "diff")

narwhals/_pandas_like/series.py

@@ -735,9 +735,8 @@ def cum_sum(self: Self, *, reverse: bool) -> Self:
         )
         return self._from_native_series(result)
 
-    def unique(self, *, maintain_order: bool = False) -> PandasLikeSeries:
-        # The param `maintain_order` is only here for compatibility with the Polars API
-        # and has no effect on the output.
+    def unique(self) -> PandasLikeSeries:
+        # pandas seems to always maintain order - is that guaranteed?
         return self._from_native_series(
             self._native_series.__class__(
                 self._native_series.unique(), name=self._native_series.name

narwhals/stable/v1/__init__.py

@@ -935,6 +935,8 @@ def unique(self, *, maintain_order: bool | None = None) -> Self:
         return self.__class__(
             lambda plx: self._to_compliant_expr(plx).unique(),
             self._is_order_dependent,
+            changes_length=True,
+            aggregates=self._aggregates,
         )
 
     def sort(self, *, descending: bool = False, nulls_last: bool = False) -> Self:

tests/conftest.py

@@ -205,7 +205,9 @@ def pytest_generate_tests(metafunc: pytest.Metafunc) -> None:
             *iter(LAZY_CONSTRUCTORS.keys()),
         ]
         selected_constructors = [
-            x for x in selected_constructors if x not in GPU_CONSTRUCTORS
+            x
+            for x in selected_constructors
+            if x not in GPU_CONSTRUCTORS and x not in "modin"  # too slow
         ]
     else:  # pragma: no cover
         selected_constructors = metafunc.config.getoption("constructors").split(",")
@@ -242,7 +244,7 @@ def pytest_generate_tests(metafunc: pytest.Metafunc) -> None:
         if (
             any(
                 x in str(metafunc.module)
-                for x in ("list", "unpivot", "from_dict", "from_numpy", "tail")
+                for x in ("list", "unpivot", "from_dict", "from_numpy")
             )
             and LAZY_CONSTRUCTORS["duckdb"] in constructors
         ):