BUG: astype(errors=ignore) for EAs (#56350)

doc/source/whatsnew/v2.2.0.rst

@@ -557,6 +557,7 @@ Numeric
 Conversion
 ^^^^^^^^^^
 - Bug in :meth:`DataFrame.astype` when called with ``str`` on unpickled array - the array might change in-place (:issue:`54654`)
+- Bug in :meth:`DataFrame.astype` where ``errors="ignore"`` had no effect for extension types (:issue:`54654`)
 - Bug in :meth:`Series.convert_dtypes` not converting all NA column to ``null[pyarrow]`` (:issue:`55346`)
 -
 

pandas/core/generic.py

@@ -6659,7 +6659,9 @@ def astype(
                 return self.copy(deep=copy)
             # GH 18099/22869: columnwise conversion to extension dtype
             # GH 24704: self.items handles duplicate column names
-            results = [ser.astype(dtype, copy=copy) for _, ser in self.items()]
+            results = [
+                ser.astype(dtype, copy=copy, errors=errors) for _, ser in self.items()
+            ]
 
         else:
             # else, only a single dtype is given

pandas/tests/extension/test_arrow.py

@@ -1475,6 +1475,13 @@ def test_astype_float_from_non_pyarrow_str():
     tm.assert_series_equal(result, expected)
 
 
+def test_astype_errors_ignore():
+    # GH 55399
+    expected = pd.DataFrame({"col": [17000000]}, dtype="int32[pyarrow]")
+    result = expected.astype("float[pyarrow]", errors="ignore")
+    tm.assert_frame_equal(result, expected)
+
+
 def test_to_numpy_with_defaults(data):
     # GH49973
     result = data.to_numpy()