Bug mixing DMatrix's with and without feature names

python-package/xgboost/core.py

@@ -513,6 +513,9 @@ class DMatrix(object):
         -------
         feature_names : list or None
         """
+        if self._feature_names is None:
+            return ['f{0}'.format(i) for i in range(self.num_col())]
+        else:
             return self._feature_names
 
     @property
@@ -567,7 +570,7 @@ class DMatrix(object):
         """
         if feature_types is not None:
 
-            if self.feature_names is None:
+            if self._feature_names is None:
                 msg = 'Unable to set feature types before setting names'
                 raise ValueError(msg)
 

python-package/xgboost/sklearn.py

@@ -497,17 +497,8 @@ class XGBClassifier(XGBModel, XGBClassifierBase):
         """
         b = self.booster()
         fs = b.get_fscore()
-        if b.feature_names is None:
-            keys = [int(k.replace('f', '')) for k in fs.keys()]
-            all_features_dict = dict.fromkeys(range(0, self._features_count), 0)
-            fs_dict = dict(zip(keys, fs.values()))
-            all_features_dict.update(fs_dict)
-            all_features = np.fromiter(all_features_dict.values(),
-                                       dtype=np.float32)
-        else:
         all_features = [fs.get(f, 0.) for f in b.feature_names]
         all_features = np.array(all_features, dtype=np.float32)
-
         return all_features / all_features.sum()
 
 

tests/python/test_basic.py

@@ -91,7 +91,7 @@ class TestBasic(unittest.TestCase):
 
         # reset
         dm.feature_names = None
-        assert dm.feature_names is None
+        self.assertEqual(dm.feature_names, ['f0', 'f1', 'f2', 'f3', 'f4'])
         assert dm.feature_types is None
 
     def test_feature_names(self):

tests/python/test_basic_models.py

@@ -99,3 +99,20 @@ class TestModels(unittest.TestCase):
         num_round = 2
         xgb.cv(param, dtrain, num_round, nfold=5,
                metrics={'error'}, seed=0, show_stdv=False)
+
+    def test_feature_names_validation(self):
+        X = np.random.random((10, 3))
+        y = np.random.randint(2, size=(10,))
+
+        dm1 = xgb.DMatrix(X, y)
+        dm2 = xgb.DMatrix(X, y, feature_names=("a", "b", "c"))
+
+        bst = xgb.train([], dm1)
+        bst.predict(dm1)  # success
+        self.assertRaises(ValueError, bst.predict, dm2)
+        bst.predict(dm1)  # success
+
+        bst = xgb.train([], dm2)
+        bst.predict(dm2)  # success
+        self.assertRaises(ValueError, bst.predict, dm1)
+        bst.predict(dm2)  # success