FIX-#6968: Align API with pandas

environment-dev.yml

@@ -5,7 +5,8 @@ dependencies:
   - pip
 
   # required dependencies
-  - pandas>=2.2,<2.3
+  # excluding 2.2.0 because of https://github.com/modin-project/modin/issues/6968
+  - pandas>=2.2,<2.3,!=2.2.0
   - numpy>=1.22.4
   - fsspec>=2022.11.0
   - packaging>=21.0

modin/pandas/groupby.py

@@ -682,14 +682,34 @@
             )
         )
 
-    def first(self, numeric_only=False, min_count=-1):
+    def first(self, numeric_only=False, min_count=-1, skipna=True):
+        # TODO: check if it works properly if we simply pass 'skipna' to
+        # the query compiler level
+        if not skipna:
+            return self._default_to_pandas(
+                lambda df: df.first(
+                    numeric_only=numeric_only,
+                    min_count=min_count,
+                    skipna=skipna,
+                )
+            )
         return self._wrap_aggregation(
             type(self._query_compiler).groupby_first,
             agg_kwargs=dict(min_count=min_count),
             numeric_only=numeric_only,
         )
 
-    def last(self, numeric_only=False, min_count=-1):
+    def last(self, numeric_only=False, min_count=-1, skipna=True):
+        # TODO: check if it works properly if we simply pass 'skipna' to
+        # the query compiler level
+        if not skipna:
+            return self._default_to_pandas(
+                lambda df: df.first(
+                    numeric_only=numeric_only,
+                    min_count=min_count,
+                    skipna=skipna,
+                )
+            )
         return self._wrap_aggregation(
             type(self._query_compiler).groupby_last,
             agg_kwargs=dict(min_count=min_count),

modin/pandas/series.py

@@ -706,13 +706,19 @@ def argmin(self, axis=None, skipna=True, *args, **kwargs):  # noqa: PR01, RT01,
             result = -1
         return result
 
-    def argsort(self, axis=0, kind="quicksort", order=None):  # noqa: PR01, RT01, D200
+    def argsort(
+        self, axis=0, kind="quicksort", order=None, stable=None
+    ):  # noqa: PR01, RT01, D200
         """
         Return the integer indices that would sort the Series values.
         """
         return self.__constructor__(
             query_compiler=self._query_compiler.argsort(
-                axis=axis, kind=kind, order=order
+                # 'stable' parameter has no effect in Pandas and is only accepted
+                # for compatibility with NumPy, so we're not passing it forward on purpose
+                axis=axis,
+                kind=kind,
+                order=order,
             )
         )
 

requirements-dev.txt

@@ -1,5 +1,6 @@
 ## required dependencies
-pandas>=2.2,<2.3
+# excluding 2.2.0 because of https://github.com/modin-project/modin/issues/6968
+pandas>=2.2,<2.3,!=2.2.0
 numpy>=1.22.4
 fsspec>=2022.11.0
 packaging>=21.0

requirements/env_hdk.yml

@@ -5,7 +5,7 @@ dependencies:
   - pip
 
   # required dependencies
-  - pandas>=2.2,<2.3
+  - pandas>=2.2,<2.3,!=2.2.0
   - numpy>=1.22.4
   - pyhdk==0.9
   - fsspec>=2022.11.0

requirements/env_unidist_linux.yml

@@ -5,7 +5,8 @@ dependencies:
   - pip
 
   # required dependencies
-  - pandas>=2.2,<2.3
+  # excluding 2.2.0 because of https://github.com/modin-project/modin/issues/6968
+  - pandas>=2.2,<2.3,!=2.2.0
   - numpy>=1.22.4
   - unidist-mpi>=0.2.1
   - mpich

requirements/env_unidist_win.yml

@@ -5,7 +5,8 @@ dependencies:
   - pip
 
   # required dependencies
-  - pandas>=2.2,<2.3
+  # excluding 2.2.0 because of https://github.com/modin-project/modin/issues/6968
+  - pandas>=2.2,<2.3,!=2.2.0
   - numpy>=1.22.4
   - unidist-mpi>=0.2.1
   - msmpi

requirements/requirements-no-engine.yml

@@ -4,7 +4,8 @@ dependencies:
   - pip
 
   # required dependencies
-  - pandas>=2.2,<2.3
+  # excluding 2.2.0 because of https://github.com/modin-project/modin/issues/6968
+  - pandas>=2.2,<2.3,!=2.2.0
   - numpy>=1.22.4
   - fsspec>=2022.11.0
   - packaging>=21.0

setup.py

@@ -51,7 +51,8 @@ def make_distribution(self):
     long_description=long_description,
     long_description_content_type="text/markdown",
     install_requires=[
-        "pandas>=2.2,<2.3",
+        # excluding 2.2.0 because of https://github.com/modin-project/modin/issues/6968
+        "pandas>=2.2,<2.3,!=2.2.0",
         "packaging>=21.0",
         "numpy>=1.22.4",
         "fsspec>=2022.11.0",