add int8_t and int16_t to join_t, make sure correct index is returned…

… for categorical joins

pandas/_libs/join.pyx

@@ -242,6 +242,8 @@ ctypedef fused join_t:
     float64_t
     float32_t
     object
+    int8_t
+    int16_t
     int32_t
     int64_t
     uint64_t

pandas/core/indexes/base.py

@@ -3822,6 +3822,9 @@ def _join_monotonic(self, other, how="left", return_indexers=False):
                 join_index, lidx, ridx = self._outer_indexer(sv, ov)
             join_index = self._wrap_joined_index(join_index, other)
 
+        if self._typ == "categoricalindex":
+            join_index = self._create_from_codes(join_index)
+
         if return_indexers:
             lidx = None if lidx is None else ensure_platform_int(lidx)
             ridx = None if ridx is None else ensure_platform_int(ridx)

pandas/tests/reshape/merge/test_merge.py

@@ -2163,3 +2163,17 @@ def test_merge_datetime_upcast_dtype():
         }
     )
     tm.assert_frame_equal(result, expected)
+
+
+def test_categorical_non_unique_monotonic():
+    # GH 28189
+    df = DataFrame(range(4), columns=["value"], index=CategoricalIndex(["1"] * 4))
+    df2 = DataFrame([[6]], columns=["value"], index=CategoricalIndex(["1"]))
+
+    result = merge(df, df2, how="left", left_index=True, right_index=True)
+    expected = DataFrame(
+        [[0, 6], [1, 6], [2, 6], [3, 6]],
+        columns=["value_x", "value_y"],
+        index=CategoricalIndex(["1"] * 4),
+    )
+    tm.assert_frame_equal(expected, result)