fix: Answer online requests that carry zero entity rows

Three empty-input shapes reached internal exceptions, and the feature server turned each into an HTTP 500: entity_rows=[] -> IndexError: list index out of range entities={"driver_id": []} -> KeyError: Missing join key values for keys: [] entities={} -> KeyError: 'pop from an empty set' The middle case is a well-formed request. The join key was supplied, it simply has no values, which is what a caller sends when the upstream query matched nothing that run. get_online_features built its columnar dict from entity_rows[0] on both the sync and async paths, so an empty list raised before anything else ran. _validate_entity_values ended with set_of_row_lengths.pop(); for a mapping with no columns the set is empty and pop() raised. Report zero rows instead. _get_unique_entities treated a join key present with zero values the same as a join key never supplied, even though the row-wise conversion just below it already returns empty results for that case. Raise only when nothing at all was supplied for the view, which leaves the existing behaviour intact: a partially supplied key set still proceeds, and a caller that supplied nothing relevant still errors with the same message listing the expected keys. That error is now MissingJoinKeyValuesException, carrying HTTP 400 so the server reports a client error rather than a 500. It subclasses KeyError as well, since that is what this condition raised before and callers catch it. Signed-off-by: Daksha1611 <mehtadaksha1611@gmail.com>
Daksha1611 · Sep 7, 2026 · Sep 8, 2026 · Sep 11, 2026 · Oct 6, 2026 · Sep 9, 2026
commit 49d6c5581c7517c5f140c5aaeb199cce732b3573
@@ -507,6 +507,24 @@ def __init__(self):
         )
 
 
+class MissingJoinKeyValuesException(FeastError, KeyError):
+    """A required join key was not supplied for the requested feature views.
+
+    Subclasses ``KeyError`` as well, because that is what this condition raised
+    before it carried an HTTP status, and callers catch it.
+    """
+
+    def __init__(self, missing_keys, empty_keys, provided_keys):
+        super().__init__(
+            f"Missing join key values for keys: {sorted(missing_keys)}. "
+            f"No values provided for keys: {sorted(empty_keys)}. "
+            f"Provided join_key_values: {list(provided_keys)}"
+        )
+
+    def http_status_code(self) -> int:
+        return HttpStatusCode.HTTP_400_BAD_REQUEST
+
+
 class PushSourceNotFoundException(FeastError):
     def __init__(self, push_source_name: str):
         super().__init__(f"Unable to find push source '{push_source_name}'.")

@@ -166,7 +166,10 @@ def get_online_features(
         include_feature_view_version_metadata: bool = False,
     ) -> OnlineResponse:
         if isinstance(entity_rows, list):
-            columnar: Dict[str, List[Any]] = {k: [] for k in entity_rows[0].keys()}
+            # An empty batch carries no column names; downstream handles zero rows.
+            columnar: Dict[str, List[Any]] = (
+                {k: [] for k in entity_rows[0].keys()} if entity_rows else {}
+            )
             for entity_row in entity_rows:
                 for key, value in entity_row.items():
                     try:
@@ -450,7 +453,10 @@ async def get_online_features_async(
         include_feature_view_version_metadata: bool = False,
     ) -> OnlineResponse:
         if isinstance(entity_rows, list):
-            columnar: Dict[str, List[Any]] = {k: [] for k in entity_rows[0].keys()}
+            # An empty batch carries no column names; downstream handles zero rows.
+            columnar: Dict[str, List[Any]] = (
+                {k: [] for k in entity_rows[0].keys()} if entity_rows else {}
+            )
             for entity_row in entity_rows:
                 for key, value in entity_row.items():
                     try:

@@ -33,6 +33,7 @@
 from feast.errors import (
     FeatureNameCollisionError,
     FeatureViewNotFoundException,
+    MissingJoinKeyValuesException,
     RequestDataNotFoundInEntityRowsException,
 )
 from feast.field import Field
@@ -556,7 +557,8 @@ def _validate_entity_values(join_key_values: Dict[str, List[ValueProto]]):
     set_of_row_lengths = {len(v) for v in join_key_values.values()}
     if len(set_of_row_lengths) > 1:
         raise ValueError("All entity rows must have the same columns.")
-    return set_of_row_lengths.pop()
+    # No columns at all means no rows; popping an empty set would raise here.
+    return set_of_row_lengths.pop() if set_of_row_lengths else 0
 
 
 def _validate_feature_refs(feature_refs: List[str], full_feature_names: bool = False):
@@ -1025,11 +1027,13 @@ def _get_unique_entities(
     )
 
     if missing_keys or empty_keys:
-        if not any(table_entity_values.values()):
-            raise KeyError(
-                f"Missing join key values for keys: {missing_keys}. "
-                f"No values provided for keys: {empty_keys}. "
-                f"Provided join_key_values: {list(join_key_values.keys())}"
+        # Columns present but all empty is a well-formed request that simply has
+        # no rows this run; the row-wise conversion below already returns empty
+        # results for it. Only a caller that supplied nothing at all for this
+        # view is reporting an error.
+        if not table_entity_values:
+            raise MissingJoinKeyValuesException(
+                missing_keys, empty_keys, join_key_values.keys()
             )
 
     # Convert the column-oriented table_entity_values into row-wise data.

@@ -0,0 +1,54 @@
+"""An empty batch of entities must not surface internal errors.
+
+A caller whose upstream query matched nothing sends zero rows. That used to hit
+``IndexError: list index out of range`` or ``KeyError: 'pop from an empty set'``,
+which the feature server turned into a 500.
+"""
+
+import pytest
+from fastapi import status as HttpStatusCode
+
+from feast.errors import MissingJoinKeyValuesException
+from feast.utils import _validate_entity_values
+from tests.utils.cli_repo_creator import CliRunner, get_example_repo
+
+FEATURES = ["driver_locations:lat", "driver_locations:lon"]
+
+
+def test_validate_entity_values_treats_no_columns_as_zero_rows():
+    """Popping the empty set raised instead of reporting zero rows."""
+    assert _validate_entity_values({}) == 0
+
+
+def test_missing_join_key_values_is_a_client_error():
+    exc = MissingJoinKeyValuesException(["driver_id"], [], [])
+    # Callers caught KeyError before this carried a status; keep that working.
+    assert isinstance(exc, KeyError)
+    assert exc.http_status_code() == HttpStatusCode.HTTP_400_BAD_REQUEST
+
+
+def test_join_key_present_with_zero_values_returns_an_empty_response():
+    """A well-formed request that simply has no rows this run."""
+    runner = CliRunner()
+    with runner.local_repo(
+        get_example_repo("example_feature_repo_1.py"), "file"
+    ) as store:
+        response = store.get_online_features(
+            features=FEATURES, entity_rows={"driver_id": []}
+        ).to_dict()
+
+    assert set(response.keys()) == {"driver_id", "lat", "lon"}
+    assert all(values == [] for values in response.values())
+
+
+@pytest.mark.parametrize("entity_rows", [[], {}], ids=["empty_list", "no_columns"])
+def test_request_without_join_keys_raises_a_client_error(entity_rows):
+    """No column names at all: report the missing join key, not an internal error."""
+    runner = CliRunner()
+    with runner.local_repo(
+        get_example_repo("example_feature_repo_1.py"), "file"
+    ) as store:
+        with pytest.raises(MissingJoinKeyValuesException) as excinfo:
+            store.get_online_features(features=FEATURES, entity_rows=entity_rows)
+
+    assert "driver_id" in str(excinfo.value)