Pin Polars version <1.35

conda/environments/all_cuda-129_arch-aarch64.yaml

@@ -66,7 +66,7 @@ dependencies:
 - pandas
 - pandas>=2.0,<2.4.0dev0
 - pandoc
-- polars>=1.29,<1.34
+- polars>=1.29,<1.35
 - pre-commit
 - pyarrow>=15.0.0
 - pydata-sphinx-theme>=0.15.4

conda/environments/all_cuda-129_arch-x86_64.yaml

@@ -67,7 +67,7 @@ dependencies:
 - pandas
 - pandas>=2.0,<2.4.0dev0
 - pandoc
-- polars>=1.29,<1.34
+- polars>=1.29,<1.35
 - pre-commit
 - pyarrow>=15.0.0
 - pydata-sphinx-theme>=0.15.4

conda/environments/all_cuda-130_arch-aarch64.yaml

@@ -66,7 +66,7 @@ dependencies:
 - pandas
 - pandas>=2.0,<2.4.0dev0
 - pandoc
-- polars>=1.29,<1.34
+- polars>=1.29,<1.35
 - pre-commit
 - pyarrow>=15.0.0
 - pydata-sphinx-theme>=0.15.4

conda/environments/all_cuda-130_arch-x86_64.yaml

@@ -67,7 +67,7 @@ dependencies:
 - pandas
 - pandas>=2.0,<2.4.0dev0
 - pandoc
-- polars>=1.29,<1.34
+- polars>=1.29,<1.35
 - pre-commit
 - pyarrow>=15.0.0
 - pydata-sphinx-theme>=0.15.4

conda/recipes/cudf-polars/recipe.yaml

@@ -52,7 +52,7 @@ requirements:
     - nvidia-ml-py>=12
     - python
     - pylibcudf =${{ version }}
-    - polars >=1.29,<1.34
+    - polars >=1.29,<1.35
     - ${{ pin_compatible("cuda-version", upper_bound="x", lower_bound="x") }}
     - if: python == "3.10"
       then: typing_extensions

dependencies.yaml

@@ -738,7 +738,7 @@ dependencies:
           # 'nvidia-ml-py' provides the 'pynvml' module
           - &nvidia_ml_py nvidia-ml-py>=12
           - packaging
-          - polars>=1.29,<1.34
+          - polars>=1.29,<1.35
     specific:
       - output_types: [requirements, pyproject]
         matrices:

python/cudf_polars/cudf_polars/containers/dataframe.py

@@ -393,7 +393,7 @@ def rename_columns(self, mapping: Mapping[str, str]) -> Self:
             stream=self.stream,
         )
 
-    def select_columns(self, names: Set[str]) -> list[Column]:
+    def select_columns(self, names: Iterable[str]) -> list[Column]:
         """Select columns by name."""
         return [c for c in self.columns if c.name in names]
 

python/cudf_polars/cudf_polars/dsl/expressions/aggregation.py

@@ -9,22 +9,30 @@
 from functools import partial
 from typing import TYPE_CHECKING, Any, ClassVar
 
+import polars as pl
+
 import pylibcudf as plc
 
-from cudf_polars.containers import Column
+from cudf_polars.containers import Column, DataType
 from cudf_polars.dsl.expressions.base import ExecutionContext, Expr
 from cudf_polars.dsl.expressions.literal import Literal
 
 if TYPE_CHECKING:
     from rmm.pylibrmm.stream import Stream
 
-    from cudf_polars.containers import DataFrame, DataType
+    from cudf_polars.containers import DataFrame
 
 __all__ = ["Agg"]
 
 
 class Agg(Expr):
-    __slots__ = ("context", "name", "op", "options", "request")
+    __slots__ = (
+        "context",
+        "name",
+        "op",
+        "options",
+        "request",
+    )
     _non_child = ("dtype", "name", "options", "context")
 
     def __init__(
@@ -156,22 +164,44 @@ def agg_request(self) -> plc.aggregation.Aggregation:  # noqa: D102
     def _reduce(
         self, column: Column, *, request: plc.aggregation.Aggregation, stream: Stream
     ) -> Column:
-        if (
-            self.name in {"mean", "median"}
-            and plc.traits.is_fixed_point(column.dtype.plc_type)
-            and self.dtype.plc_type.id() in {plc.TypeId.FLOAT32, plc.TypeId.FLOAT64}
+        is_mean_or_median = self.name in {"mean", "median"}
+        is_quantile = self.name == "quantile"
+
+        out_dtype = self.dtype
+        if plc.traits.is_fixed_point(column.dtype.plc_type) and (
+            is_mean_or_median or is_quantile
         ):
-            column = column.astype(self.dtype, stream=stream)
+            cast_to = (
+                self.dtype
+                if is_mean_or_median
+                and plc.traits.is_floating_point(self.dtype.plc_type)
+                else DataType(pl.Float64())
+            )
+            column = column.astype(cast_to, stream=stream)
+            out_dtype = cast_to
+        if column.size == 0 or column.null_count == column.size:
+            res = None
+            if self.name == "n_unique":
+                res = 0 if column.size == 0 else 1
+            return Column(
+                plc.Column.from_scalar(
+                    plc.Scalar.from_py(res, out_dtype.plc_type, stream=stream),
+                    1,
+                    stream=stream,
+                ),
+                name=column.name,
+                dtype=out_dtype,
+            )
         return Column(
             plc.Column.from_scalar(
                 plc.reduce.reduce(
-                    column.obj, request, self.dtype.plc_type, stream=stream
+                    column.obj, request, out_dtype.plc_type, stream=stream
                 ),
                 1,
                 stream=stream,
             ),
             name=column.name,
-            dtype=self.dtype,
+            dtype=out_dtype,
         )
 
     def _count(self, column: Column, *, include_nulls: bool, stream: Stream) -> Column:
@@ -199,6 +229,21 @@ def _sum(self, column: Column, stream: Stream) -> Column:
                 name=column.name,
                 dtype=self.dtype,
             )
+        if plc.traits.is_fixed_point(column.dtype.plc_type):
+            return Column(
+                plc.Column.from_scalar(
+                    plc.reduce.reduce(
+                        column.obj,
+                        plc.aggregation.sum(),
+                        column.dtype.plc_type,
+                        stream=stream,
+                    ),
+                    1,
+                    stream=stream,
+                ),
+                name=column.name,
+                dtype=column.dtype,
+            )
         return self._reduce(column, request=plc.aggregation.sum(), stream=stream)
 
     def _min(self, column: Column, *, propagate_nans: bool, stream: Stream) -> Column:

python/cudf_polars/cudf_polars/dsl/expressions/binaryop.py

@@ -8,13 +8,15 @@
 
 from typing import TYPE_CHECKING, ClassVar
 
-from polars.polars import _expr_nodes as pl_expr
+from polars import polars
 
 import pylibcudf as plc
 
 from cudf_polars.containers import Column
 from cudf_polars.dsl.expressions.base import ExecutionContext, Expr
 
+pl_expr = polars._expr_nodes
+
 if TYPE_CHECKING:
     from cudf_polars.containers import DataFrame, DataType
 
@@ -59,7 +61,9 @@ def __init__(
         plc.binaryop.BinaryOperator.LOGICAL_OR: plc.binaryop.BinaryOperator.NULL_LOGICAL_OR,
     }
 
-    _MAPPING: ClassVar[dict[pl_expr.Operator, plc.binaryop.BinaryOperator]] = {
+    _MAPPING: ClassVar[
+        dict[polars._expr_nodes.Operator, plc.binaryop.BinaryOperator]
+    ] = {
         pl_expr.Operator.Eq: plc.binaryop.BinaryOperator.EQUAL,
         pl_expr.Operator.EqValidity: plc.binaryop.BinaryOperator.NULL_EQUALS,
         pl_expr.Operator.NotEq: plc.binaryop.BinaryOperator.NOT_EQUAL,

python/cudf_polars/cudf_polars/dsl/expressions/boolean.py

@@ -22,7 +22,7 @@
     from typing_extensions import Self
 
     import polars.type_aliases as pl_types
-    from polars.polars import _expr_nodes as pl_expr
+    from polars import polars
 
     from rmm.pylibrmm.stream import Stream
 
@@ -55,7 +55,7 @@ class Name(IntEnum):
         Not = auto()
 
         @classmethod
-        def from_polars(cls, obj: pl_expr.BooleanFunction) -> Self:
+        def from_polars(cls, obj: polars._expr_nodes.BooleanFunction) -> Self:
             """Convert from polars' `BooleanFunction`."""
             try:
                 function, name = str(obj).split(".", maxsplit=1)

python/cudf_polars/cudf_polars/dsl/expressions/datetime.py

@@ -17,7 +17,7 @@
 if TYPE_CHECKING:
     from typing_extensions import Self
 
-    from polars.polars import _expr_nodes as pl_expr
+    from polars import polars
 
     from cudf_polars.containers import DataFrame, DataType
 
@@ -75,7 +75,7 @@ class Name(IntEnum):
         Year = auto()
 
         @classmethod
-        def from_polars(cls, obj: pl_expr.TemporalFunction) -> Self:
+        def from_polars(cls, obj: polars._expr_nodes.TemporalFunction) -> Self:
             """Convert from polars' `TemporalFunction`."""
             try:
                 function, name = str(obj).split(".", maxsplit=1)

python/cudf_polars/cudf_polars/dsl/expressions/string.py

@@ -12,8 +12,8 @@
 from enum import IntEnum, auto
 from typing import TYPE_CHECKING, Any, ClassVar
 
+from polars import polars
 from polars.exceptions import InvalidOperationError
-from polars.polars import dtype_str_repr
 
 import pylibcudf as plc
 
@@ -26,8 +26,6 @@
 if TYPE_CHECKING:
     from typing_extensions import Self
 
-    from polars.polars import _expr_nodes as pl_expr
-
     from cudf_polars.containers import DataFrame, DataType
 
 __all__ = ["StringFunction"]
@@ -101,7 +99,7 @@ class Name(IntEnum):
         ZFill = auto()
 
         @classmethod
-        def from_polars(cls, obj: pl_expr.StringFunction) -> Self:
+        def from_polars(cls, obj: polars._expr_nodes.StringFunction) -> Self:
             """Convert from polars' `StringFunction`."""
             try:
                 function, name = str(obj).split(".", maxsplit=1)
@@ -283,7 +281,7 @@ def _validate_input(self) -> None:
                     and width.value is not None
                     and width.value < 0
                 ):  # pragma: no cover
-                    dtypestr = dtype_str_repr(width.dtype.polars_type)
+                    dtypestr = polars.dtype_str_repr(width.dtype.polars_type)
                     raise InvalidOperationError(
                         f"conversion from `{dtypestr}` to `u64` "
                         f"failed in column 'literal' for 1 out of "

python/cudf_polars/cudf_polars/dsl/expressions/struct.py

@@ -18,7 +18,7 @@
 if TYPE_CHECKING:
     from typing_extensions import Self
 
-    from polars.polars import _expr_nodes as pl_expr
+    from polars import polars
 
     from cudf_polars.containers import DataFrame, DataType
 
@@ -42,7 +42,7 @@ class Name(IntEnum):
         )  # https://github.com/pola-rs/polars/pull/23022#issuecomment-2933910958
 
         @classmethod
-        def from_polars(cls, obj: pl_expr.StructFunction) -> Self:
+        def from_polars(cls, obj: polars._expr_nodes.StructFunction) -> Self:
             """Convert from polars' `StructFunction`."""
             try:
                 function, name = str(obj).split(".", maxsplit=1)

python/cudf_polars/cudf_polars/dsl/expressions/unary.py

@@ -150,7 +150,7 @@ def __init__(
         )
 
         if self.name not in UnaryFunction._supported_fns:
-            raise NotImplementedError(f"Unary function {name=}")
+            raise NotImplementedError(f"Unary function {name=}")  # pragma: no cover
         if self.name in UnaryFunction._supported_cum_aggs:
             (reverse,) = self.options
             if reverse: