Oracle Machine Learning for R provides datastores that you can use to save OML4R proxy objects, as well as any R object, in an Oracle database.

You can grant or revoke read privilege access to a datastore for one or more users. You can restore the saved objects in another R session. The objects in a datastore are also accessible to Embedded R Execution through both the R and the SQL interfaces.

This section describes the OML4R functions that you can use to create and manage datastores. The section contains the following topics:

• About Persisting Oracle Machine Learning for R Objects
  With OML4R datastores, you can save R objects in the database.
• About OML4R Datastores
  Each database schema has a table that stores named OML4R datastores.
• Save Objects to a Datastore
  The ore.save function saves one or more R objects in the specified datastore.
• Control Access to Datastores
  With the ore.grant and ore.revoke functions you can grant or revoke access to an OML4R datastore.
• Get Information about Datastore Contents
  You can get information about a datastore in the current user schema by using the ore.datastore and ore.datastoreSummary functions.
• Restore Objects from a Datastore
  The ore.load function restores R objects saved in a datastore to the R global environment, .GlobalEnv.
• Delete a Datastore
  With the ore.delete function, you can delete objects from an OML4R datastore or you can delete the datastore itself.
• About Using a Datastore in Embedded R Execution
  Saving objects in a datastore makes it very easy to pass arguments to, and reference R objects with, Embedded R Execution functions.