feast-dev · ntkathole · Oct 1, 2026 · Sep 30, 2026
@@ -883,7 +883,12 @@ def _augment_response_with_on_demand_transforms(
     """
     from feast.online_response import OnlineResponse
 
-    requested_odfv_map = {odfv.name: odfv for odfv in requested_on_demand_feature_views}
+    # Feature refs name an ODFV by its alias when it was added to a
+    # FeatureService with ``with_name``, so key the lookup the same way.
+    requested_odfv_map = {
+        (odfv.projection.name_alias if odfv.projection else None) or odfv.name: odfv
+        for odfv in requested_on_demand_feature_views
+    }
     requested_odfv_feature_names = requested_odfv_map.keys()
 
     odfv_feature_refs = defaultdict(list)

@@ -8,6 +8,7 @@
 
 from feast import (
     Entity,
+    FeatureService,
     FeatureStore,
     FeatureView,
     FileSource,
@@ -597,3 +598,60 @@ def aliased_guard(inputs: pd.DataFrame) -> pd.DataFrame:
         # the aliased-but-declared source must survive; the unrelated FV stays hidden
         assert response["saw_declared"] == [True]
         assert response["saw_undeclared"] == [False]
+
+
+@pytest.mark.parametrize("mode", ["pandas", "python"])
+def test_feature_service_serves_aliased_odfv_features(mode):
+    """An ODFV added to a FeatureService under ``with_name`` must still be
+    computed. Its feature refs use the alias, so the transform lookup has to be
+    keyed by the alias too, not only by the ODFV's registered name."""
+    with tempfile.TemporaryDirectory() as data_dir:
+        store, driver, src, fv1, fv2, driver_df = _two_fv_store(data_dir)
+
+        if mode == "pandas":
+
+            @on_demand_feature_view(
+                sources=[fv1],
+                schema=[Field(name="conv_rate_x10", dtype=Float64)],
+                mode="pandas",
+            )
+            def scaled(inputs: pd.DataFrame) -> pd.DataFrame:
+                out = pd.DataFrame()
+                out["conv_rate_x10"] = inputs["conv_rate"] * 10
+                return out
+
+        else:
+
+            @on_demand_feature_view(
+                sources=[fv1],
+                schema=[Field(name="conv_rate_x10", dtype=Float64)],
+                mode="python",
+            )
+            def scaled(inputs: dict) -> dict:
+                return {"conv_rate_x10": [v * 10 for v in inputs["conv_rate"]]}
+
+        service = FeatureService(
+            name="aliased_service",
+            features=[fv1, scaled.with_name("scaled_alias")],
+        )
+        store.apply([driver, src, fv1, scaled, service])
+        store.write_to_online_store(feature_view_name="fv1", df=driver_df)
+
+        response = store.get_online_features(
+            entity_rows=[{"driver_id": 1001}],
+            features=store.get_feature_service("aliased_service"),
+        ).to_dict()
+        assert response["conv_rate_x10"] == [
+            pytest.approx(response["conv_rate"][0] * 10)
+        ]
+
+        if mode != "pandas":
+            return
+        response = store.get_online_features(
+            entity_rows=[{"driver_id": 1001}],
+            features=store.get_feature_service("aliased_service"),
+            full_feature_names=True,
+        ).to_dict()
+        assert response["scaled_alias__conv_rate_x10"] == [
+            pytest.approx(response["fv1__conv_rate"][0] * 10)
+        ]