◐ Shell
clean mode source ↗

fix: Fix STRING type handling in on-demand feature views with RequestSource by Copilot · Pull Request #5657 · feast-dev/feast

Copilot

Added 'object' dtype mapping to STRING in type_map.py since pandas
uses object dtype for string columns by default. Added test case to
verify STRING type inputs work with RequestSource.

Co-authored-by: franciscojavierarceo <4163062+franciscojavierarceo@users.noreply.github.com>
Enhanced the fix to check actual value types when type_name='object'.
This prevents incorrect type inference for mixed-type object columns
while still fixing the STRING type issue. Falls back to STRING when
value is None or unknown type.

Co-authored-by: franciscojavierarceo <4163062+franciscojavierarceo@users.noreply.github.com>

Copilot AI changed the title [WIP] Fix ondemand feature view apply failure for STRING inputs Fix STRING type handling in on-demand feature views with RequestSource

Oct 11, 2025 
Moved object dtype special handling after array type checks to prevent
intercepting numpy arrays with object dtype. Added check to exclude
lists/arrays from object dtype value-based inference. This fixes unit
test failures where arrays with object dtype were incorrectly mapped
to STRING instead of being processed as arrays.

Co-authored-by: franciscojavierarceo <4163062+franciscojavierarceo@users.noreply.github.com>

@franciscojavierarceo franciscojavierarceo changed the title Fix STRING type handling in on-demand feature views with RequestSource fix: Fix STRING type handling in on-demand feature views with RequestSource

Oct 12, 2025 
The object dtype check was returning STRING early even when the value
was an array/list, preventing the array handling logic from running.
Moved the array check to the outer condition so arrays with object
dtype fall through to the list/array handling code which correctly
returns STRING_LIST.

Co-authored-by: franciscojavierarceo <4163062+franciscojavierarceo@users.noreply.github.com>