Feature groups can be made using python and the procedure is quite similar to the one that was followed for adding a SQL Feature Group. Let's say you created a project under Personalized Recommendations use case and uploaded the datasets, "Movies Ratings" and "Movies Metadata". Now, let's say you need to combine these two datasets for creating a new feature group using python. You can follow the steps below to do so:
Click on the "Add New Feature Group" button and then on "Python feature group" to create the feature group:
A configuration window be opened. Enter the name for the feature group and python function in "Table Name" and "Function Name" fields. It's optional to add a "Description" for the feature group so you can leave it empty. Now, select the appropriate feature groups in "Input Feature Groups" which in our case would be "movie_ratings" and "movies_metadata". Next, select the "CPU Size" and "Memory" in GBs to finalize the computing power and memory you wish to assign to execute the python code. Finally, write the python function to create the feature group. In our case, the code merges both Input Feature Groups on "movie_id" to create a new feature group:
Now, click on the "Save changes to Feature Group" button to create the feature group. The feature group will be created with the provided configuration. You can also preview the results by materializing the latest feature group version and then clicking on the "View" button for your version under "Feature Group Versions" list:
