1.4.2.1 About Oracle Machine Learning for R Data Types and Classes

OML4R has data types that map R data types to SQL data types.

In an R session, when you create database objects from R objects or you create R objects from database data, OML4R translates R data types to SQL data types and the reverse where possible.

OML4R creates objects that are instances of OML4R classes. OML4R overloads many standard R functions so that they use OML4R classes and data types. R language constructs and syntax are supported for objects that are mapped to Oracle Database objects.

Table 1-1 Mappings Between R, OML4R, and SQL Data Types

R Data Type	OML4R Data Type	SQL Data Type
character mode vector	ore.character	VARCHAR2 INTERVAL YEAR TO MONTH
integer mode vector	ore.integer	NUMBER
logical mode vector	ore.logical	The NUMBER 0 for FALSE and 1 for TRUE
numeric mode vector	ore.number	BINARY_DOUBLE BINARY_FLOAT FLOAT NUMBER
Date	ore.date	DATE
POSIXct POSIXlt	ore.datetime	TIMESTAMP TIMESTAMP WITH TIME ZONE TIMESTAMP WITH LOCAL TIME ZONE
difftime	ore.difftime	INTERVAL DAY TO SECOND
None	Not supported	LONG LONG RAW RAW User defined data types Reference data types

Note:

• Objects of type ore.datetime do not support a time zone setting, instead they use the system time zone Sys.timezone if it is available or GMT if Sys.timezone is not available.

• The SQL VARCHAR2 data type is mapped to the R character data type through the embedded R input data argument. Users can convert the character variable to a factor in R if needed by using as.factor().