SNOW-1852934: Add support for Dataframe.map

snowflakedb · Dec 13, 2024 · 7c0fa6e · 7c0fa6e
1 parent 71843f3
commit 7c0fa6e
Show file tree

Hide file tree

Showing 4 changed files with 28 additions and 15 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -16,6 +16,7 @@
 - Added support for `Series.str.center`.
 - Added support for `Series.str.pad`.
 - Added support for applying Snowpark Python function `snowflake_cortex_sentiment`.
+- Added support for `DataFrame.map`.
 
 
 ## 1.26.0 (2024-12-05)

diff --git a/docs/source/modin/supported/dataframe_supported.rst b/docs/source/modin/supported/dataframe_supported.rst
@@ -258,7 +258,7 @@ Methods
 +-----------------------------+---------------------------------+----------------------------------+----------------------------------------------------+
 | ``lt``                      | P                               | ``level``                        |                                                    |
 +-----------------------------+---------------------------------+----------------------------------+----------------------------------------------------+
-| ``map``                     | N                               |                                  |                                                    |
+| ``map``                     | P                               |                                  | ``N`` if ``na_action == "ignore"``                 |
 +-----------------------------+---------------------------------+----------------------------------+----------------------------------------------------+
 | ``mask``                    | P                               |                                  | ``N`` if given ``axis`` when ``other`` is a        |
 |                             |                                 |                                  | ``DataFrame`` or ``level`` parameters;             |

diff --git a/src/snowflake/snowpark/modin/plugin/extensions/dataframe_overrides.py b/src/snowflake/snowpark/modin/plugin/extensions/dataframe_overrides.py
@@ -137,9 +137,15 @@ def decorator(base_method: Any):
 
 # Avoid overwriting builtin `map` by accident
 @register_dataframe_accessor("map")
-@dataframe_not_implemented()
-def _map(self, func, na_action: str | None = None, **kwargs) -> DataFrame:
-    pass  # pragma: no cover
+def _map(self, func: PythonFuncType, na_action: str | None = None, **kwargs):
+    # TODO: SNOW-1063346: Modin upgrade - modin.pandas.DataFrame functions
+    if not callable(func):
+        raise TypeError(f"{func} is not callable")
+    return self.__constructor__(
+        query_compiler=self._query_compiler.applymap(
+            func, na_action=na_action, **kwargs
+        )
+    )
 
 
 @register_dataframe_not_implemented()
@@ -804,14 +810,12 @@ def apply(
 # Snowpark pandas uses a separate QC method, while modin directly calls map.
 @register_dataframe_accessor("applymap")
 def applymap(self, func: PythonFuncType, na_action: str | None = None, **kwargs):
-    # TODO: SNOW-1063346: Modin upgrade - modin.pandas.DataFrame functions
-    if not callable(func):
-        raise TypeError(f"{func} is not callable")
-    return self.__constructor__(
-        query_compiler=self._query_compiler.applymap(
-            func, na_action=na_action, **kwargs
-        )
+    warnings.warn(
+        "DataFrame.applymap has been deprecated. Use DataFrame.map instead.",
+        FutureWarning,
+        stacklevel=2,
     )
+    return self.map(func, na_action=na_action, **kwargs)
 
 
 # We need to override _get_columns to satisfy

diff --git a/tests/integ/modin/frame/test_applymap.py b/tests/integ/modin/frame/test_applymap.py
@@ -28,25 +28,33 @@
 )
 
 
+@pytest.fixture(params=["applymap", "map"])
+def method(request):
+    """
+    how keyword to pass to merge.
+    """
+    return request.param
+
+
 @pytest.mark.parametrize("data,func,return_type", BASIC_DATA_FUNC_RETURN_TYPE_MAP)
 @sql_count_checker(query_count=7, udf_count=1)
-def test_applymap_basic_without_type_hints(data, func, return_type):
+def test_applymap_basic_without_type_hints(data, func, return_type, method):
     frame_data = {0: data, 1: data}
     native_df = native_pd.DataFrame(frame_data)
     snow_df = pd.DataFrame(frame_data)
-    eval_snowpark_pandas_result(snow_df, native_df, lambda x: x.applymap(func))
+    eval_snowpark_pandas_result(snow_df, native_df, lambda x: getattr(x, method)(func))
 
 
 @pytest.mark.parametrize("data,func,return_type", BASIC_DATA_FUNC_RETURN_TYPE_MAP)
 @sql_count_checker(query_count=7, udf_count=1)
-def test_applymap_basic_with_type_hints(data, func, return_type):
+def test_applymap_basic_with_type_hints(data, func, return_type, method):
     func_with_type_hint = create_func_with_return_type_hint(func, return_type)
 
     frame_data = {0: data, 1: data}
     native_df = native_pd.DataFrame(frame_data)
     snow_df = pd.DataFrame(frame_data)
     eval_snowpark_pandas_result(
-        snow_df, native_df, lambda x: x.applymap(func_with_type_hint)
+        snow_df, native_df, lambda x: getattr(x, method)(func_with_type_hint)
     )