Closed Ben-Epstein closed 3 years ago
Support returning training sets as pandas dataframes dataframes (under 50MB)
This enables users to use the featurestore without spark and get small datasets as pandas dataframes
More flexibility with the feature store usage, no need for spark on small datasets.
ml-workflow
Against K8s
Support training sets as JSOn and Pandas (under 50MB)
The return_sql parameter is deprecated and will be removed in a future release
return_sql
Support returning training sets as pandas dataframes dataframes (under 50MB)
Description
This enables users to use the featurestore without spark and get small datasets as pandas dataframes
Motivation and Context
More flexibility with the feature store usage, no need for spark on small datasets.
Dependencies
ml-workflow
How Has This Been Tested?
Against K8s
Screenshots (if appropriate):
Changelog Inclusions
Additions
Support training sets as JSOn and Pandas (under 50MB)
Changes
Fixes
Deprecated
The
return_sqlparameter is deprecated and will be removed in a future releaseRemoved
Breaking Changes