1. User's Guide
  2. Get Started with Oracle Machine Learning for R
  3. Use OML4R with On-Premises Oracle Database
  4. Create and Manage R Objects in Oracle Database
  5. Create Temporary Database Tables

2.2.3.3 Create Temporary Database Tables

You can create temporary database tables, and their corresponding proxy ore.frame objects, from local R data.frame objects with the ore.push function.

With the ore.pull function you can create a local R object that contains a copy of data represented by an OML4R proxy object.

The ore.push function translates an R object into an OML4R object of the appropriate data type. The ore.pull function takes an ore class object and returns an R object. If the input object is an ore.list, the ore.pull function creates a data.frame and translates the data of each database column into the appropriate R representation.

Note:

You can pull data to a local R data.frame only if the data can fit into the R session memory. Also, even if the data fits in memory but is still very large, you may not be able to perform many, or any, R functions in the client R session.

Unless you explicitly save the proxy objects in a datastore in the database that were created using the ore.push function, the temporary tables and proxy objects are discarded when you quit your R session.

Example 2-14 Using ore.push and ore.pull to Move Data

This example demonstrates pushing an R data.frame object to the database as a temporary database table with an associated ore.frame object, iris_of, then creating another ore.frame object, iris_of_setosa, by selecting one column from iris_of, and then pulling the iris_of_setosa object into the local R session memory as a data.frame object. The example displays the class of some of the objects. 

class(iris)
# Push the iris data frame to the database.
iris_of <- ore.push(iris)
class(iris_of)
# Display the data type of the Sepal.Length column in the data.frame.
class(iris$Sepal.Length)
# Display the data type of the Sepal.Length column in the ore.frame.
class(iris_of$Sepal.Length)
# Filter one column of the data set.
iris_of_setosa <- iris_of[iris_of$Species == "setosa", ]
class(iris_of_setosa)
# Pull the selected column into the local R memory.
local_setosa = ore.pull(iris_of_setosa)
class(local_setosa)
Listing for This Example
R> class(iris)
[1] "data.frame"
R> # Push the iris data frame to the database.
R> iris_of <- ore.push(iris)
R> class(iris_of)
[1] "ore.frame"
attr(,"package")
[1] "OREbase"
R> # Display the data type of the Sepal.Length column in the data.frame.
R> class(iris$Sepal.Length)
[1] "numeric"
R> # Display the data type of the Sepal.Length column in the ore.frame.
R> class(iris_of$Sepal.Length)
[1] "ore.numeric"
attr(,"package")
[1] "OREbase"
R> # Filter one column of the data set.
R> iris_of_setosa <- iris_of[iris_of$Species == "setosa", ]
R> class(iris_of_setosa)
[1] "ore.frame"
attr(,"package")
[1] "OREbase"
R> # Pull the selected column into the local R memory.
R> local_setosa = ore.pull(iris_of_setosa)
R> class(local_setosa)
[1] "data.frame"

Parent topic: Create and Manage R Objects in Oracle Database