thomasjpfan
diff --git a/‎doc/whats_new/v0.24.rst
Copy file name to clipboardExpand all lines: doc/whats_new/v0.24.rst
+3Lines changed: 3 additions & 0 deletions b/‎doc/whats_new/v0.24.rst
Copy file name to clipboardExpand all lines: doc/whats_new/v0.24.rst
+3Lines changed: 3 additions & 0 deletions
diff --git a/‎doc/whats_new/v1.0.rst
Copy file name to clipboardExpand all lines: doc/whats_new/v1.0.rst
+1-1Lines changed: 1 addition & 1 deletion b/‎doc/whats_new/v1.0.rst
Copy file name to clipboardExpand all lines: doc/whats_new/v1.0.rst
+1-1Lines changed: 1 addition & 1 deletion
diff --git a/‎sklearn/preprocessing/_encoders.py
Copy file name to clipboardExpand all lines: sklearn/preprocessing/_encoders.py
+6Lines changed: 6 additions & 0 deletions b/‎sklearn/preprocessing/_encoders.py
Copy file name to clipboardExpand all lines: sklearn/preprocessing/_encoders.py
+6Lines changed: 6 additions & 0 deletions
diff --git a/‎sklearn/preprocessing/tests/test_encoders.py
Copy file name to clipboardExpand all lines: sklearn/preprocessing/tests/test_encoders.py
+20Lines changed: 20 additions & 0 deletions b/‎sklearn/preprocessing/tests/test_encoders.py
Copy file name to clipboardExpand all lines: sklearn/preprocessing/tests/test_encoders.py
+20Lines changed: 20 additions & 0 deletions
@@ -53,6 +53,9 @@ Changelog
   `'use_encoded_value'` strategies.
   :pr:`19234` by `Guillaume Lemaitre <glemaitre>`.
 
+- |Fix| :meth:`preprocessing.OrdinalEncoder.transfrom` correctly handles
+  unknown values for string dtypes. :pr:`19888` by `Thomas Fan`_.
+
 :mod:`sklearn.multioutput`
 ..........................
 
 
@@ -311,7 +311,7 @@ Changelog
   :pr:`18649` by `Leandro Hermida <hermidalc>` and
   `Rodion Martynov <marrodion>`.
 
-- |Fix| The `fit` method of the successive halving parameter search 
+- |Fix| The `fit` method of the successive halving parameter search
   (:class:`model_selection.HalvingGridSearchCV`, and
   :class:`model_selection.HalvingRandomSearchCV`) now correctly handles the
   `groups` parameter. :pr:`19847` by :user:`Xiaoyu Chai <xiaoyuchai>`.
 
@@ -167,6 +167,12 @@ def _transform(self, X, handle_unknown='error', force_all_finite=True,
                     if (self.categories_[i].dtype.kind in ('U', 'S')
                             and self.categories_[i].itemsize > Xi.itemsize):
                         Xi = Xi.astype(self.categories_[i].dtype)
+                    elif (self.categories_[i].dtype.kind == 'O' and
+                            Xi.dtype.kind == 'U'):
+                        # categories are objects and Xi are numpy strings.
+                        # Cast Xi to an object dtype to prevent truncation
+                        # when setting invalid values.
+                        Xi = Xi.astype('O')
                     else:
                         Xi = Xi.copy()
 
 
@@ -1580,3 +1580,23 @@ def test_ordinal_encoder_sparse():
     X_trans_sparse = sparse.csr_matrix(X_trans)
     with pytest.raises(TypeError, match=err_msg):
         encoder.inverse_transform(X_trans_sparse)
+
+
+@pytest.mark.parametrize("X_train", [
+    [['AA', 'B']],
+    np.array([['AA', 'B']], dtype='O'),
+    np.array([['AA', 'B']], dtype='U'),
+])
+@pytest.mark.parametrize("X_test", [
+    [['A', 'B']],
+    np.array([['A', 'B']], dtype='O'),
+    np.array([['A', 'B']], dtype='U'),
+])
+def test_ordinal_encoder_handle_unknown_string_dtypes(X_train, X_test):
+    """Checks that ordinal encoder transforms string dtypes. Non-regression
+    test for #19872."""
+    enc = OrdinalEncoder(handle_unknown='use_encoded_value', unknown_value=-9)
+    enc.fit(X_train)
+
+    X_trans = enc.transform(X_test)
+    assert_allclose(X_trans, [[-9, 0]])