groupby -> group_by update

Author: MarcoGorelli

check_completeness.py

@@ -36,16 +36,16 @@
     if i not in polars_spec:
         print(f"Series.{i} missing from polars spec!")
 
-# groupby
+# group_by
 spec = dataframe_api.GroupBy().__dir__()
 pandas_spec = (
     pandas_standard.PandasDataFrame(pd.DataFrame({"a": [1]}), api_version="2023.08-beta")
-    .groupby(["a"])
+    .group_by(["a"])
     .__dir__()
 )
 polars_spec = (
     polars_standard.PolarsDataFrame(pl.DataFrame({"a": [1]}), api_version="2023.08-beta")
-    .groupby(["a"])
+    .group_by(["a"])
     .__dir__()
 )
 
@@ -61,7 +61,7 @@
     "Sequence",
     "column_object",
     "dataframe_object",
-    "groupby_object",
+    "group_by_object",
     "DType",
 }
 spec = [i for i in dataframe_api.__dir__() if i not in exclude and not i.startswith("_")]

dataframe_api_compat/pandas_standard/pandas_standard.py

@@ -373,7 +373,7 @@ def _validate_result(self, result: pd.DataFrame) -> None:
             # defensive check
             raise AssertionError(
                 "Groupby operation could not be performed on columns "
-                f"{failed_columns}. Please drop them before calling groupby."
+                f"{failed_columns}. Please drop them before calling group_by."
             )
 
     def size(self) -> PandasDataFrame:
@@ -511,7 +511,7 @@ def dataframe(self) -> pd.DataFrame:
     def shape(self) -> tuple[int, int]:
         return self.dataframe.shape
 
-    def groupby(self, keys: Sequence[str]) -> PandasGroupBy:
+    def group_by(self, keys: Sequence[str]) -> PandasGroupBy:
         if not isinstance(keys, collections.abc.Sequence):
             raise TypeError(f"Expected sequence of strings, got: {type(keys)}")
         if isinstance(keys, str):
@@ -606,6 +606,9 @@ def rename_columns(self, mapping: Mapping[str, str]) -> PandasDataFrame:
         )
 
     def get_column_names(self) -> list[str]:
+        # DO NOT REMOVE
+        # This one is used in upstream tests - even if deprecated,
+        # just leave it in for backwards compatibility
         return self.dataframe.columns.tolist()
 
     def sorted_indices(

dataframe_api_compat/polars_standard/polars_standard.py

@@ -770,7 +770,9 @@ def __init__(self, df: pl.LazyFrame, keys: Sequence[str], api_version: str) -> N
         self.df = df
         self.keys = keys
         self._api_version = api_version
-        self.group_by = self.df.groupby if pl.__version__ < "0.19.0" else self.df.group_by
+        self.group_by = (
+            self.df.group_by if pl.__version__ < "0.19.0" else self.df.group_by
+        )
 
     def size(self) -> PolarsDataFrame:
         result = self.group_by(self.keys).count().rename({"count": "size"})
@@ -866,7 +868,7 @@ def __dataframe_namespace__(self) -> Any:
     def dataframe(self) -> pl.LazyFrame:
         return self.df
 
-    def groupby(self, keys: Sequence[str]) -> PolarsGroupBy:
+    def group_by(self, keys: Sequence[str]) -> PolarsGroupBy:
         return PolarsGroupBy(self.df, keys, api_version=self._api_version)
 
     def get_column_by_name(self, name: str) -> PolarsColumn[DType]:
@@ -938,6 +940,9 @@ def rename_columns(self, mapping: Mapping[str, str]) -> PolarsDataFrame:
         )
 
     def get_column_names(self) -> list[str]:
+        # DO NOT REMOVE
+        # This one is used in upstream tests - even if deprecated,
+        # just leave it in for backwards compatibility
         return self.dataframe.columns
 
     def __eq__(  # type: ignore[override]

tests/groupby/groupby_any_all_test.py

@@ -17,15 +17,15 @@
         ("all", [False, True], [False, False]),
     ],
 )
-def test_groupby_boolean(
+def test_group_by_boolean(
     library: str,
     aggregation: str,
     expected_b: list[bool],
     expected_c: list[bool],
     request: pytest.FixtureRequest,
 ) -> None:
     df = bool_dataframe_2(library)
-    result = getattr(df.groupby(["key"]), aggregation)()
+    result = getattr(df.group_by(["key"]), aggregation)()
     # need to sort
     idx = result.sorted_indices(["key"])
     result = result.get_rows(idx)
@@ -35,9 +35,9 @@ def test_groupby_boolean(
     pd.testing.assert_frame_equal(result_pd, expected)
 
 
-def test_groupby_invalid_any_all(library: str, request: pytest.FixtureRequest) -> None:
+def test_group_by_invalid_any_all(library: str, request: pytest.FixtureRequest) -> None:
     df = integer_dataframe_4(library)
     with pytest.raises((ValueError, SchemaError)):
-        df.groupby(["key"]).any()
+        df.group_by(["key"]).any()
     with pytest.raises((ValueError, SchemaError)):
-        df.groupby(["key"]).all()
+        df.group_by(["key"]).all()

tests/groupby/invalid_test.py

@@ -5,11 +5,11 @@
 from tests.utils import integer_dataframe_1
 
 
-def test_groupby_invalid(library: str) -> None:
+def test_group_by_invalid(library: str) -> None:
     df = integer_dataframe_1(library).select(["a"])
     with pytest.raises((KeyError, TypeError)):
-        df.groupby(0)
+        df.group_by(0)
     with pytest.raises((KeyError, TypeError)):
-        df.groupby("0")
+        df.group_by("0")
     with pytest.raises((KeyError, TypeError)):
-        df.groupby(["b"])
+        df.group_by(["b"])

tests/groupby/numeric_test.py

@@ -25,15 +25,15 @@
         ("var", [0.5, 0.5], [0.5, 0.5]),
     ],
 )
-def test_groupby_numeric(
+def test_group_by_numeric(
     library: str,
     aggregation: str,
     expected_b: list[float],
     expected_c: list[float],
     request: pytest.FixtureRequest,
 ) -> None:
     df = integer_dataframe_4(library)
-    result = getattr(df.groupby(["key"]), aggregation)()
+    result = getattr(df.group_by(["key"]), aggregation)()
     sorted_indices = result.sorted_indices(["key"])
     result = result.get_rows(sorted_indices)
     result_pd = interchange_to_pandas(result, library)

tests/groupby/size_test.py

@@ -12,9 +12,9 @@
     import pytest
 
 
-def test_groupby_size(library: str, request: pytest.FixtureRequest) -> None:
+def test_group_by_size(library: str, request: pytest.FixtureRequest) -> None:
     df = integer_dataframe_4(library)
-    result = df.groupby(["key"]).size()
+    result = df.group_by(["key"]).size()
     # got to sort
     idx = result.sorted_indices(["key"])
     result = result.get_rows(idx)