fix find

Author: Rohan Jain

pandas/core/arrays/arrow/array.py

@@ -2328,19 +2328,25 @@ def _str_fullmatch(
         return self._str_match(pat, case, flags, na)
 
     def _str_find(self, sub: str, start: int = 0, end: int | None = None) -> Self:
-        if start != 0 and end is not None:
-            slices = pc.utf8_slice_codeunits(self._pa_array, start, stop=end)
-            result = pc.find_substring(slices, sub)
-            not_found = pc.equal(result, -1)
-            start_offset = max(0, start)
-            offset_result = pc.add(result, start_offset)
-            result = pc.if_else(not_found, result, offset_result)
-        elif start == 0 and end is None:
-            slices = self._pa_array
-            result = pc.find_substring(slices, sub)
+        if (start == 0 or start is None) and end is None:
+            result = pc.find_substring(self._pa_array, sub)
         else:
-            raise NotImplementedError(
-                f"find not implemented with {sub=}, {start=}, {end=}"
+            result = pc.find_substring(self._pa_array, sub)
+            length = pc.utf8_length(self._pa_array)
+            if start is None:
+                start = pa.scalar(0, result.type)
+            elif start < 0:
+                start = pc.add(start, length)
+            if end is None:
+                end = length
+            elif end < 0:
+                end = pc.add(end, length)
+            found = pc.not_equal(pa.scalar(-1, type=result.type), result)
+            found_in_bounds = pc.and_(
+                pc.greater_equal(result, start), pc.less(result, end)
+            )
+            result = pc.if_else(
+                pc.and_(found, found_in_bounds), result, pa.scalar(-1, type=result.type)
             )
         return type(self)(result)
 

pandas/tests/extension/test_arrow.py

@@ -1919,28 +1919,30 @@ def test_str_fullmatch(pat, case, na, exp):
 
 
 @pytest.mark.parametrize(
-    "sub, start, end, exp, exp_typ",
-    [["ab", 0, None, [0, None], pa.int32()], ["bc", 1, 3, [1, None], pa.int64()]],
+    "sub, start, end, exp",
+    [["ab", 0, None, [0, None]], ["bc", 1, 3, [1, None]], ["ab", 1, None, [-1, None]]],
 )
-def test_str_find(sub, start, end, exp, exp_typ):
+def test_str_find(sub, start, end, exp):
     ser = pd.Series(["abc", None], dtype=ArrowDtype(pa.string()))
     result = ser.str.find(sub, start=start, end=end)
-    expected = pd.Series(exp, dtype=ArrowDtype(exp_typ))
+    expected = pd.Series(exp, dtype=ArrowDtype(pa.int32()))
     tm.assert_series_equal(result, expected)
 
 
 def test_str_find_negative_start():
     # GH 56411
     ser = pd.Series(["abc", None], dtype=ArrowDtype(pa.string()))
     result = ser.str.find(sub="b", start=-1000, end=3)
-    expected = pd.Series([1, None], dtype=ArrowDtype(pa.int64()))
+    expected = pd.Series([1, None], dtype=ArrowDtype(pa.int32()))
     tm.assert_series_equal(result, expected)
 
 
-def test_str_find_notimplemented():
-    ser = pd.Series(["abc", None], dtype=ArrowDtype(pa.string()))
-    with pytest.raises(NotImplementedError, match="find not implemented"):
-        ser.str.find("ab", start=1)
+def test_str_find_negative_start_negative_end():
+    # GH 56411
+    ser = pd.Series(["abcdefg", None], dtype=ArrowDtype(pa.string()))
+    result = ser.str.find(sub="d", start=-6, end=-3)
+    expected = pd.Series([3, None], dtype=ArrowDtype(pa.int32()))
+    tm.assert_series_equal(result, expected)
 
 
 @pytest.mark.parametrize(