BUG(string dtype): Empty sum produces incorrect result

doc/source/whatsnew/v2.3.0.rst

@@ -107,6 +107,7 @@ Conversion
 
 Strings
 ^^^^^^^
+- Bug in :meth:`DataFrame.sum` with ``axis=1``, :meth:`.DataFrameGroupBy.sum` or :meth:`.SeriesGroupBy.sum` with ``skipna=True``, and :meth:`.Resampler.sum` on :class:`StringDtype` with all NA values resulted in ``0`` and is now the empty string ``""`` (:issue:`60229`)
 - Bug in :meth:`Series.__pos__` and :meth:`DataFrame.__pos__` did not raise for :class:`StringDtype` with ``storage="pyarrow"`` (:issue:`60710`)
 - Bug in :meth:`Series.rank` for :class:`StringDtype` with ``storage="pyarrow"`` incorrectly returning integer results in case of ``method="average"`` and raising an error if it would truncate results (:issue:`59768`)
 - Bug in :meth:`Series.replace` with :class:`StringDtype` when replacing with a non-string value was not upcasting to ``object`` dtype (:issue:`60282`)

pandas/core/arrays/base.py

@@ -2628,7 +2628,15 @@ def _groupby_op(
             if op.how not in ["any", "all"]:
                 # Fail early to avoid conversion to object
                 op._get_cython_function(op.kind, op.how, np.dtype(object), False)
-            npvalues = self.to_numpy(object, na_value=np.nan)
+
+            arr = self
+            if op.how == "sum":
+                # https://github.com/pandas-dev/pandas/issues/60229
+                # All NA should result in the empty string.
+                assert "skipna" in kwargs
+                if kwargs["skipna"] and min_count == 0:
+                    arr = arr.fillna("")
+            npvalues = arr.to_numpy(object, na_value=np.nan)
         else:
             raise NotImplementedError(
                 f"function is not implemented for this dtype: {self.dtype}"

pandas/tests/frame/test_reductions.py

@@ -835,6 +835,16 @@ def test_axis_1_empty(self, all_reductions, index):
         expected = Series([], index=index, dtype=expected_dtype)
         tm.assert_series_equal(result, expected)
 
+    @pytest.mark.parametrize("min_count", [0, 1])
+    def test_axis_1_sum_na(self, string_dtype_no_object, skipna, min_count):
+        # https://github.com/pandas-dev/pandas/issues/60229
+        dtype = string_dtype_no_object
+        df = DataFrame({"a": [pd.NA]}, dtype=dtype)
+        result = df.sum(axis=1, skipna=skipna, min_count=min_count)
+        value = "" if skipna and min_count == 0 else pd.NA
+        expected = Series([value], dtype=dtype)
+        tm.assert_series_equal(result, expected)
+
     @pytest.mark.parametrize("method, unit", [("sum", 0), ("prod", 1)])
     @pytest.mark.parametrize("numeric_only", [None, True, False])
     def test_sum_prod_nanops(self, method, unit, numeric_only):

pandas/tests/groupby/test_reductions.py

@@ -955,6 +955,20 @@ def test_min_empty_string_dtype(func, string_dtype_no_object):
     tm.assert_frame_equal(result, expected)
 
 
+@pytest.mark.parametrize("min_count", [0, 1])
+def test_string_dtype_empty_sum(string_dtype_no_object, skipna, min_count):
+    # https://github.com/pandas-dev/pandas/issues/60229
+    dtype = string_dtype_no_object
+    df = DataFrame({"a": ["x"], "b": [pd.NA]}, dtype=dtype)
+    gb = df.groupby("a")
+    result = gb.sum(skipna=skipna, min_count=min_count)
+    value = "" if skipna and min_count == 0 else pd.NA
+    expected = DataFrame(
+        {"b": value}, index=pd.Index(["x"], name="a", dtype=dtype), dtype=dtype
+    )
+    tm.assert_frame_equal(result, expected)
+
+
 def test_max_nan_bug():
     df = DataFrame(
         {

pandas/tests/resample/test_base.py

@@ -223,6 +223,31 @@ def test_resample_empty_series(freq, index, resample_method):
     assert result.index.freq == expected.index.freq
 
 
+@pytest.mark.parametrize("min_count", [0, 1])
+def test_resample_empty_sum_string(string_dtype_no_object, min_count):
+    # https://github.com/pandas-dev/pandas/issues/60229
+    dtype = string_dtype_no_object
+    ser = Series(
+        pd.NA,
+        index=DatetimeIndex(
+            [
+                "2000-01-01 00:00:00",
+                "2000-01-01 00:00:10",
+                "2000-01-01 00:00:20",
+                "2000-01-01 00:00:30",
+            ]
+        ),
+        dtype=dtype,
+    )
+    rs = ser.resample("20s")
+    result = rs.sum(min_count=min_count)
+
+    value = "" if min_count == 0 else pd.NA
+    index = date_range(start="2000-01-01", freq="20s", periods=2, unit="s")
+    expected = Series(value, index=index, dtype=dtype)
+    tm.assert_series_equal(result, expected)
+
+
 @pytest.mark.parametrize(
     "freq",
     [

pandas/tests/resample/test_resampler_grouper.py

@@ -494,6 +494,26 @@ def test_resample_groupby_agg_object_dtype_all_nan(consolidate):
     tm.assert_frame_equal(result, expected)
 
 
+@pytest.mark.parametrize("min_count", [0, 1])
+def test_groupby_resample_empty_sum_string(
+    string_dtype_no_object, test_frame, min_count
+):
+    # https://github.com/pandas-dev/pandas/issues/60229
+    dtype = string_dtype_no_object
+    test_frame = test_frame.assign(B=pd.array([pd.NA] * len(test_frame), dtype=dtype))
+    gbrs = test_frame.groupby("A").resample("40s")
+    result = gbrs.sum(min_count=min_count)
+
+    index = pd.MultiIndex(
+        levels=[[1, 2, 3], [pd.to_datetime("2000-01-01", unit="ns")]],
+        codes=[[0, 1, 2], [0, 0, 0]],
+        names=["A", None],
+    )
+    value = "" if min_count == 0 else pd.NA
+    expected = DataFrame({"B": value}, index=index, dtype=dtype)
+    tm.assert_frame_equal(result, expected)
+
+
 def test_groupby_resample_with_list_of_keys():
     # GH 47362
     df = DataFrame(