BUG: pd.eval with engine="numexpr" fails with float division

doc/source/whatsnew/v3.0.0.rst

@@ -698,6 +698,7 @@ Other
 - Bug in :class:`DataFrame` when passing a ``dict`` with a NA scalar and ``columns`` that would always return ``np.nan`` (:issue:`57205`)
 - Bug in :func:`eval` on :class:`ExtensionArray` on including division ``/`` failed with a ``TypeError``. (:issue:`58748`)
 - Bug in :func:`eval` where the names of the :class:`Series` were not preserved when using ``engine="numexpr"``. (:issue:`10239`)
+- Bug in :func:`eval` with ``engine="numexpr"`` returning unexpected result for float division. (:issue:`59736`)
 - Bug in :func:`unique` on :class:`Index` not always returning :class:`Index` (:issue:`57043`)
 - Bug in :meth:`DataFrame.apply` where passing ``engine="numba"`` ignored ``args`` passed to the applied function (:issue:`58712`)
 - Bug in :meth:`DataFrame.eval` and :meth:`DataFrame.query` which caused an exception when using NumPy attributes via ``@`` notation, e.g., ``df.eval("@np.floor(a)")``. (:issue:`58041`)

pandas/core/computation/align.py

@@ -213,7 +213,7 @@ def reconstruct_object(typ, obj, axes, dtype, name):
     if hasattr(res_t, "type") and typ == np.bool_ and res_t != np.bool_:
         ret_value = res_t.type(obj)
     else:
-        ret_value = typ(obj).astype(res_t)
+        ret_value = res_t.type(obj)
         # The condition is to distinguish 0-dim array (returned in case of
         # scalar) and 1 element array
         # e.g. np.array(0) and np.array([0])

pandas/tests/computation/test_eval.py

@@ -1998,3 +1998,11 @@ def test_validate_bool_args(value):
     msg = 'For argument "inplace" expected type bool, received type'
     with pytest.raises(ValueError, match=msg):
         pd.eval("2+2", inplace=value)
+
+
+@td.skip_if_no("numexpr")
+def test_eval_float_div_numexpr():
+    # GH 59736
+    result = pd.eval("1 / 2", engine="numexpr")
+    expected = 0.5
+    assert result == expected