refactor(python): Remove re-export of data type groups

py-polars/polars/__init__.py

@@ -1,5 +1,3 @@
-from __future__ import annotations
-
 import contextlib
 import os
 
@@ -40,13 +38,6 @@
 )
 from polars.dataframe import DataFrame
 from polars.datatypes import (
-    DATETIME_DTYPES,
-    DURATION_DTYPES,
-    FLOAT_DTYPES,
-    INTEGER_DTYPES,
-    NESTED_DTYPES,
-    NUMERIC_DTYPES,
-    TEMPORAL_DTYPES,
     Array,
     Binary,
     Boolean,
@@ -250,14 +241,6 @@
     "UInt64",
     "Unknown",
     "Utf8",
-    # polars.datatypes: dtype groups
-    "DATETIME_DTYPES",
-    "DURATION_DTYPES",
-    "FLOAT_DTYPES",
-    "INTEGER_DTYPES",
-    "NESTED_DTYPES",
-    "NUMERIC_DTYPES",
-    "TEMPORAL_DTYPES",
     # polars.io
     "read_avro",
     "read_clipboard",
@@ -401,7 +384,7 @@
 os.environ["POLARS_ALLOW_EXTENSION"] = "true"
 
 
-def __getattr__(name: str) -> type[Exception]:
+def __getattr__(name: str):  # type: ignore[no-untyped-def]
     # Deprecate re-export of exceptions at top-level
     if name in dir(exceptions):
         from polars._utils.deprecation import issue_deprecation_warning
@@ -416,5 +399,20 @@ def __getattr__(name: str) -> type[Exception]:
         )
         return getattr(exceptions, name)
 
+    # Deprecate data type groups at top-level
+    import polars.datatypes.group as dtgroup
+
+    if name in dir(dtgroup):
+        from polars._utils.deprecation import issue_deprecation_warning
+
+        issue_deprecation_warning(
+            message=(
+                f"`{name}` is deprecated. Define your own data type groups or use the"
+                " `polars.selectors` module for selecting columns of a certain data type."
+            ),
+            version="1.0.0",
+        )
+        return getattr(dtgroup, name)
+
     msg = f"module {__name__!r} has no attribute {name!r}"
     raise AttributeError(msg)

py-polars/polars/_utils/various.py

@@ -23,8 +23,6 @@
 import polars as pl
 from polars import functions as F
 from polars.datatypes import (
-    FLOAT_DTYPES,
-    INTEGER_DTYPES,
     Boolean,
     Date,
     Datetime,
@@ -34,6 +32,7 @@
     String,
     Time,
 )
+from polars.datatypes.group import FLOAT_DTYPES, INTEGER_DTYPES
 from polars.dependencies import _check_for_numpy
 from polars.dependencies import numpy as np
 

py-polars/polars/dataframe/frame.py

@@ -61,7 +61,6 @@
 from polars.dataframe._html import NotebookFormatter
 from polars.dataframe.group_by import DynamicGroupBy, GroupBy, RollingGroupBy
 from polars.datatypes import (
-    INTEGER_DTYPES,
     N_INFER_DEFAULT,
     Boolean,
     Float32,
@@ -75,6 +74,7 @@
     UInt32,
     UInt64,
 )
+from polars.datatypes.group import INTEGER_DTYPES
 from polars.dependencies import (
     _GREAT_TABLES_AVAILABLE,
     _HVPLOT_AVAILABLE,
@@ -2769,9 +2769,7 @@ def write_excel(
         dtype_formats : dict
             A `{dtype:str,}` dictionary that sets the default Excel format for the
             given dtype. (This can be overridden on a per-column basis by the
-            `column_formats` param). It is also valid to use dtype groups such as
-            `pl.FLOAT_DTYPES` as the dtype/format key, to simplify setting uniform
-            integer and float formats.
+            `column_formats` param).
         conditional_formats : dict
             A dictionary of colname (or selector) keys to a format str, dict, or list
             that defines conditional formatting options for the specified columns.
@@ -3007,7 +3005,7 @@ def write_excel(
         >>> df.write_excel(  # doctest: +SKIP
         ...     table_style="Table Style Light 2",
         ...     # apply accounting format to all flavours of integer
-        ...     dtype_formats={pl.INTEGER_DTYPES: "#,##0_);(#,##0)"},
+        ...     dtype_formats={dt: "#,##0_);(#,##0)" for dt in [pl.Int32, pl.Int64]},
         ...     sparklines={
         ...         # default options; just provide source cols
         ...         "trend": ["q1", "q2", "q3", "q4"],
@@ -8422,18 +8420,18 @@ def select(
 
         >>> with pl.Config(auto_structify=True):
         ...     df.select(
-        ...         is_odd=(pl.col(pl.INTEGER_DTYPES) % 2).name.suffix("_is_odd"),
+        ...         is_odd=(pl.col(pl.Int64) % 2 == 1).name.suffix("_is_odd"),
         ...     )
         shape: (3, 1)
-        ┌───────────┐
-        │ is_odd    │
-        │ ---       │
-        │ struct[2] │
-        ╞═══════════╡
-        │ {1,0}     │
-        │ {0,1}     │
-        │ {1,0}     │
-        └───────────┘
+        ┌──────────────┐
+        │ is_odd       │
+        │ ---          │
+        │ struct[2]    │
+        ╞══════════════╡
+        │ {true,false} │
+        │ {false,true} │
+        │ {true,false} │
+        └──────────────┘
         """
         return self.lazy().select(*exprs, **named_exprs).collect(_eager=True)
 

py-polars/polars/datatypes/__init__.py

@@ -5,7 +5,6 @@
     Categorical,
     DataType,
     DataTypeClass,
-    DataTypeGroup,
     Date,
     Datetime,
     Decimal,
@@ -34,17 +33,8 @@
     Utf8,
 )
 from polars.datatypes.constants import (
-    DATETIME_DTYPES,
     DTYPE_TEMPORAL_UNITS,
-    DURATION_DTYPES,
-    FLOAT_DTYPES,
-    INTEGER_DTYPES,
     N_INFER_DEFAULT,
-    NESTED_DTYPES,
-    NUMERIC_DTYPES,
-    SIGNED_INTEGER_DTYPES,
-    TEMPORAL_DTYPES,
-    UNSIGNED_INTEGER_DTYPES,
 )
 from polars.datatypes.constructor import (
     numpy_type_to_constructor,
@@ -72,7 +62,6 @@
     "Categorical",
     "DataType",
     "DataTypeClass",
-    "DataTypeGroup",
     "Date",
     "Datetime",
     "Decimal",
@@ -100,17 +89,8 @@
     "Unknown",
     "Utf8",
     # constants
-    "DATETIME_DTYPES",
-    "DTYPE_TEMPORAL_UNITS",
-    "DURATION_DTYPES",
-    "FLOAT_DTYPES",
-    "INTEGER_DTYPES",
-    "NESTED_DTYPES",
-    "NUMERIC_DTYPES",
     "N_INFER_DEFAULT",
-    "SIGNED_INTEGER_DTYPES",
-    "TEMPORAL_DTYPES",
-    "UNSIGNED_INTEGER_DTYPES",
+    "DTYPE_TEMPORAL_UNITS",
     # constructor
     "numpy_type_to_constructor",
     "numpy_values_and_dtype",

py-polars/polars/datatypes/classes.py

@@ -181,38 +181,6 @@ def is_nested(cls) -> bool:
         return issubclass(cls, NestedType)
 
 
-class DataTypeGroup(frozenset):  # type: ignore[type-arg]
-    """Group of data types."""
-
-    _match_base_type: bool
-
-    def __new__(
-        cls, items: Iterable[DataType | DataTypeClass], *, match_base_type: bool = True
-    ) -> DataTypeGroup:
-        """
-        Construct a DataTypeGroup.
-
-        Parameters
-        ----------
-        items :
-            iterable of data types
-        match_base_type:
-            match the base type
-        """
-        for it in items:
-            if not isinstance(it, (DataType, DataTypeClass)):
-                msg = f"DataTypeGroup items must be dtypes; found {type(it).__name__!r}"
-                raise TypeError(msg)
-        dtype_group = super().__new__(cls, items)
-        dtype_group._match_base_type = match_base_type
-        return dtype_group
-
-    def __contains__(self, item: Any) -> bool:
-        if self._match_base_type and isinstance(item, (DataType, DataTypeClass)):
-            item = item.base_type()
-        return super().__contains__(item)
-
-
 class NumericType(DataType):
     """Base class for numeric data types."""
 

py-polars/polars/datatypes/constants.py

@@ -2,85 +2,10 @@
 
 from typing import TYPE_CHECKING
 
-from polars.datatypes import (
-    Array,
-    DataTypeGroup,
-    Date,
-    Datetime,
-    Decimal,
-    Duration,
-    Float32,
-    Float64,
-    Int8,
-    Int16,
-    Int32,
-    Int64,
-    List,
-    Struct,
-    Time,
-    UInt8,
-    UInt16,
-    UInt32,
-    UInt64,
-)
-
 if TYPE_CHECKING:
-    from polars.type_aliases import (
-        PolarsDataType,
-        PolarsIntegerType,
-        PolarsTemporalType,
-        TimeUnit,
-    )
+    from polars.type_aliases import TimeUnit
 
+# Number of rows to scan by default when inferring datatypes
+N_INFER_DEFAULT = 100
 
 DTYPE_TEMPORAL_UNITS: frozenset[TimeUnit] = frozenset(["ns", "us", "ms"])
-DATETIME_DTYPES: frozenset[PolarsDataType] = DataTypeGroup(
-    [
-        Datetime,
-        Datetime("ms"),
-        Datetime("us"),
-        Datetime("ns"),
-        Datetime("ms", "*"),
-        Datetime("us", "*"),
-        Datetime("ns", "*"),
-    ]
-)
-DURATION_DTYPES: frozenset[PolarsDataType] = DataTypeGroup(
-    [
-        Duration,
-        Duration("ms"),
-        Duration("us"),
-        Duration("ns"),
-    ]
-)
-TEMPORAL_DTYPES: frozenset[PolarsTemporalType] = DataTypeGroup(
-    frozenset([Date, Time]) | DATETIME_DTYPES | DURATION_DTYPES
-)
-SIGNED_INTEGER_DTYPES: frozenset[PolarsIntegerType] = DataTypeGroup(
-    [
-        Int8,
-        Int16,
-        Int32,
-        Int64,
-    ]
-)
-UNSIGNED_INTEGER_DTYPES: frozenset[PolarsIntegerType] = DataTypeGroup(
-    [
-        UInt8,
-        UInt16,
-        UInt32,
-        UInt64,
-    ]
-)
-INTEGER_DTYPES: frozenset[PolarsIntegerType] = (
-    SIGNED_INTEGER_DTYPES | UNSIGNED_INTEGER_DTYPES
-)
-FLOAT_DTYPES: frozenset[PolarsDataType] = DataTypeGroup([Float32, Float64])
-NUMERIC_DTYPES: frozenset[PolarsDataType] = DataTypeGroup(
-    FLOAT_DTYPES | INTEGER_DTYPES | frozenset([Decimal])
-)
-
-NESTED_DTYPES: frozenset[PolarsDataType] = DataTypeGroup([List, Struct, Array])
-
-# number of rows to scan by default when inferring datatypes
-N_INFER_DEFAULT = 100