Expected Behavior

For retrieving from Redis (currently, Feast maps entities -> features), in the Java library, there is an optimization which uses HGETALL instead of HMGET when for a given entity there are more than 50 different features being retrieved.

I'd expect this to also be enforced on the Python side.

Current Behavior

The Python SDK always uses HMGET

Steps to reproduce

Specifications

• Version:
• Platform:
• Subsystem:

Possible Solution