From 224486925e5fd278963164064cc7b6d559f586e1 Mon Sep 17 00:00:00 2001
From: Owen Christie <owendot2011@gmail.com>
Date: Fri, 25 Oct 2024 00:35:38 -0400
Subject: [PATCH 1/2] Add fix for #59242

---
 doc/source/whatsnew/v2.3.0.rst        |  2 +-
 pandas/core/internals/construction.py | 12 +++++++++++-
 pandas/tests/io/test_sql.py           | 14 ++++++++++++++
 3 files changed, 26 insertions(+), 2 deletions(-)

diff --git a/doc/source/whatsnew/v2.3.0.rst b/doc/source/whatsnew/v2.3.0.rst
index 01c2ed3821d7a..2880ecec81f66 100644
--- a/doc/source/whatsnew/v2.3.0.rst
+++ b/doc/source/whatsnew/v2.3.0.rst
@@ -130,7 +130,7 @@ MultiIndex
 
 I/O
 ^^^
--
+- Bug in :func:`read_sql` causing an unintended exception when byte data was being converted to string when using the pyarrow dtype_backend (:issue:`59242`)
 -
 
 Period
diff --git a/pandas/core/internals/construction.py b/pandas/core/internals/construction.py
index 959e572b2b35b..1900ed282e876 100644
--- a/pandas/core/internals/construction.py
+++ b/pandas/core/internals/construction.py
@@ -970,7 +970,17 @@ def convert(arr):
                     if dtype_backend != "numpy" and arr.dtype == np.dtype("O"):
                         new_dtype = StringDtype()
                         arr_cls = new_dtype.construct_array_type()
-                        arr = arr_cls._from_sequence(arr, dtype=new_dtype)
+                        try:
+                            # Addressing (#59242)
+                            # Byte data that could not be decoded into
+                            # a string would throw a UnicodeDecodeError exception
+
+                            # Try and greedily convert to string
+                            # Will fail if the object is bytes
+                            arr = arr_cls._from_sequence(arr, dtype=new_dtype)
+                        except UnicodeDecodeError:
+                            pass
+
                 elif dtype_backend != "numpy" and isinstance(arr, np.ndarray):
                     if arr.dtype.kind in "iufb":
                         arr = pd_array(arr, copy=False)
diff --git a/pandas/tests/io/test_sql.py b/pandas/tests/io/test_sql.py
index c28a33069d23f..69ad44d1a5e73 100644
--- a/pandas/tests/io/test_sql.py
+++ b/pandas/tests/io/test_sql.py
@@ -4352,3 +4352,17 @@ def test_xsqlite_if_exists(sqlite_buildin):
         (5, "E"),
     ]
     drop_table(table_name, sqlite_buildin)
+
+
+def test_bytes_column(sqlite_buildin):
+    """
+    Regression test for (#59242)
+    Bytes being returned in a column that could not be converted
+    to a string would raise a UnicodeDecodeError
+    when using dtype_backend='pyarrow'
+    """
+    query = """
+    select cast(x'0123456789abcdef0123456789abcdef' as blob) a
+    """
+    df = pd.read_sql(query, sqlite_buildin, dtype_backend="pyarrow")
+    assert df.a.values[0] == b"\x01#Eg\x89\xab\xcd\xef\x01#Eg\x89\xab\xcd\xef"

From bd00fc545e25a611d97edecb9aac8c0324d17e90 Mon Sep 17 00:00:00 2001
From: Owen Christie <owendot2011@gmail.com>
Date: Fri, 25 Oct 2024 18:09:53 -0400
Subject: [PATCH 2/2] add skip import

---
 pandas/tests/io/test_sql.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/pandas/tests/io/test_sql.py b/pandas/tests/io/test_sql.py
index 69ad44d1a5e73..73f9ff42287fc 100644
--- a/pandas/tests/io/test_sql.py
+++ b/pandas/tests/io/test_sql.py
@@ -4355,6 +4355,7 @@ def test_xsqlite_if_exists(sqlite_buildin):
 
 
 def test_bytes_column(sqlite_buildin):
+    pytest.importorskip("pyarrow")
     """
     Regression test for (#59242)
     Bytes being returned in a column that could not be converted