Skip to content

fix: Support historical feature retrieval with start_date/end_date in RemoteOfflineStore #5703

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ntkathole merged 6 commits into from
Nov 3, 2025
Merged

fix: Support historical feature retrieval with start_date/end_date in RemoteOfflineStore #5703

Show file tree
Hide file tree
Changes from 4 commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
b8a8ab6
Added start_date & end_date range for historical data retrieval from …
aniketpalu Oct 21, 2025
d6bb773
Minor linting & reformatting changes
aniketpalu Oct 21, 2025
0e17fdf
Fixed minor issue of empty df with only key as column being trated as…
aniketpalu Oct 28, 2025
434ab75
Merge branch 'master' into RHOAIENG-32904
aniketpalu Nov 1, 2025
0d52d17
Removed unnecessary logic to create mock df
aniketpalu Nov 3, 2025
4bf7de7
Minor linting & formatting changes
aniketpalu Nov 3, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 10 additions & 1 deletion sdk/python/feast/infra/offline_stores/remote.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -443,7 +443,16 @@ def get_table_column_names_and_types_from_data_source(
return zip(table.column("name").to_pylist(), table.column("type").to_pylist())


def _create_retrieval_metadata(feature_refs: List[str], entity_df: pd.DataFrame):
def _create_retrieval_metadata(
feature_refs: List[str], entity_df: Optional[pd.DataFrame] = None
):
if entity_df is None:
return RetrievalMetadata(
features=feature_refs,
keys=[], # No entity keys when no entity_df provided
min_event_timestamp=None,
max_event_timestamp=None,
)
entity_schema = _get_entity_schema(
entity_df=entity_df,
)
Expand Down
15 changes: 15 additions & 0 deletions sdk/python/feast/offline_server.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,7 @@
from typing import Any, Dict, List, Optional, cast

import click
import pandas as pd
import pyarrow as pa
import pyarrow.flight as fl
from google.protobuf.json_format import Parse
Expand Down Expand Up @@ -431,6 +432,20 @@ def get_historical_features(self, command: dict, key: Optional[str] = None):
# Extract parameters from the internal flights dictionary
entity_df_value = self.flights[key]
entity_df = pa.Table.to_pandas(entity_df_value)
# Check if this is a mock/empty table (contains only 'key' column)
if len(entity_df.columns) == 1 and "key" in entity_df.columns:
entity_df = None

# If no entity_df provided, create minimal one with timestamps
if entity_df is None and "start_date" in command and "end_date" in command:
# Create minimal entity_df with event_timestamp when using start_date/end_date mode
start_date = utils.make_tzaware(
datetime.fromisoformat(command["start_date"])
)
end_date = utils.make_tzaware(datetime.fromisoformat(command["end_date"]))
entity_df = pd.DataFrame(
Copy link
Copy Markdown
Collaborator

@jyejare jyejare Nov 2, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We should not need to create dummy entity_df at all. We should handle empty(entity_df=None) entity_df for and in offline_store.get_historical_features which we are returning in this function.

Copy link
Copy Markdown
Contributor Author

@aniketpalu aniketpalu Nov 3, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You are right, its not needed to create mock dataframe with timestamps, just had to make sure that empty_df is being passed.

{"event_timestamp": pd.to_datetime([start_date, end_date])}
)

feature_view_names = command["feature_view_names"]
name_aliases = command["name_aliases"]
Expand Down
Loading