cast non-string to string for __from_arrow__

Author: jorisvandenbossche

pandas/core/arrays/string_.py

@@ -75,6 +75,10 @@
 
 from pandas.io.formats import printing
 
+if HAS_PYARROW:
+    import pyarrow as pa
+    import pyarrow.compute as pc
+
 if TYPE_CHECKING:
     from collections.abc import MutableMapping
 
@@ -337,7 +341,15 @@ def __from_arrow__(
         Construct StringArray from pyarrow Array/ChunkedArray.
         """
         if self.storage == "pyarrow":
-            from pandas.core.arrays.string_arrow import ArrowStringArray
+            from pandas.core.arrays.string_arrow import (
+                ArrowStringArray,
+                _chk_pyarrow_available,
+            )
+
+            _chk_pyarrow_available()
+
+            if not pa.types.is_large_string(array.type):
+                array = pc.cast(array, pa.large_string())
 
             return ArrowStringArray(array, dtype=self)
 

pandas/tests/io/test_parquet.py

@@ -1145,7 +1145,7 @@ def test_roundtrip_decimal(self, tmp_path, pa):
         df.to_parquet(path, schema=pa.schema([("a", pa.decimal128(5))]))
         result = read_parquet(path)
         if pa_version_under19p0:
-            expected = pd.DataFrame({"a": ["123"]}, dtype="string[python]")
+            expected = pd.DataFrame({"a": ["123"]}, dtype="string")
         else:
             expected = pd.DataFrame({"a": [Decimal("123.00")]}, dtype="object")
         tm.assert_frame_equal(result, expected)

scripts/validate_unwanted_patterns.py

@@ -58,6 +58,7 @@
     "_fill_limit_area_1d",
     "_make_block",
     "_DatetimeTZBlock",
+    "_chk_pyarrow_available",
 }