The classes in the teradataml.analytics.mle subpackage require "coprocessor nodes", the ones in the teradataml.analytics.sqle subpackage do not.
(So you can import the subpackage, if you only want to see the sqle functions, for example.)
See the Function Reference documentation link mentioned a few posts above.
Another quick question about the teradataml package. I tested a few sql features in the teradataml package and some of those are trying to create views on the teradata system. For eg: when you pass a sql to the dataframe command in teradataml, it ends up creating a view and then retreiving the data. This is an obstacle for us because users are not allowed to create views on our prod system. is there any way to make teradataml not create views on the system in order to retreive data?
teradataml provides a feature where user can specify a database where it can create the required temporary objects (the views that you are referring to.) While creating the context with create_context() API, the user can pass this database for the temp_database parameter. This will allow user to create temporary objects in the specified database rather than in default database.
create_context(host = <hostname>, username = <user>, password = <pass>, temp_database_name = <database_where_you_can_create_views>)
Let me know, if this works for you.