fix: Fix STRING type handling in on-demand feature views by ntkathole · Pull Request #5669 · feast-dev/feast
What this PR does / why we need it:
Pandas DataFrames use dtype=object for strings by default (not dtype=string), but Feast's type inference system didn't handle "object" dtype, causing it to fail when trying to convert string columns.
Which issue(s) this PR fixes:
Fixes #5641
| object_string_test_view.infer_features() | ||
|
|
||
|
|
||
| def test_on_demand_features_invalid_type_inference(): |
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
if this is invalid, shouldn't it raise?
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
it's a negative test.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There is with pytest.raises() sections which catch exceptions raised when mode is mismatched or feature is missing
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ooh nice i see on line 138, thank you.
Signed-off-by: ntkathole <nikhilkathole2683@gmail.com>