Backport PR #52606 on branch 2.0.x (BUG: to_numeric(errors='coerce', dtype_backend='pyarrow') with ArrowDtype) (#52611)

Backport PR #52606: BUG: to_numeric(errors='coerce', dtype_backend='pyarrow') with ArrowDtype

Co-authored-by: Matthew Roeschke <[email protected]>

Author: meeseeksmachine

doc/source/whatsnew/v2.0.1.rst

@@ -31,6 +31,7 @@ Bug fixes
 - Bug in :meth:`ArrowDtype.__from_arrow__` not respecting if dtype is explicitly given (:issue:`52533`)
 - Bug in :func:`read_csv` casting PyArrow datetimes to NumPy when ``dtype_backend="pyarrow"`` and ``parse_dates`` is set causing a performance bottleneck in the process (:issue:`52546`)
 - Bug in :class:`arrays.DatetimeArray` constructor returning an incorrect unit when passed a non-nanosecond numpy datetime array (:issue:`52555`)
+- Bug in :func:`to_numeric` with ``errors='coerce'`` and ``dtype_backend='pyarrow'`` with :class:`ArrowDtype` data (:issue:`52588`)
 
 .. ---------------------------------------------------------------------------
 .. _whatsnew_201.other:

pandas/core/tools/numeric.py

@@ -270,7 +270,8 @@ def to_numeric(
     # GH33013: for IntegerArray, BooleanArray & FloatingArray need to reconstruct
     # masked array
     if (mask is not None or new_mask is not None) and not is_string_dtype(values.dtype):
-        if mask is None:
+        if mask is None or (new_mask is not None and new_mask.shape == mask.shape):
+            # GH 52588
             mask = new_mask
         else:
             mask = mask.copy()

pandas/tests/tools/test_to_numeric.py

@@ -6,6 +6,7 @@
 
 import pandas as pd
 from pandas import (
+    ArrowDtype,
     DataFrame,
     Index,
     Series,
@@ -942,3 +943,12 @@ def test_invalid_dtype_backend():
     )
     with pytest.raises(ValueError, match=msg):
         to_numeric(ser, dtype_backend="numpy")
+
+
+def test_coerce_pyarrow_backend():
+    # GH 52588
+    pa = pytest.importorskip("pyarrow")
+    ser = Series(list("12x"), dtype=ArrowDtype(pa.string()))
+    result = to_numeric(ser, errors="coerce", dtype_backend="pyarrow")
+    expected = Series([1, 2, None], dtype=ArrowDtype(pa.int64()))
+    tm.assert_series_equal(result, expected)