11-04-2015 06:09 PM
How does Proc IML pass data to R? My general question is whether it'd be as fast as "embedded R" in the context of monetDB, TeraData, SQL Server, Oracle, IBM, etc. basically
Does... SAS trick R into thinking its data is an R vector, or does it force R to read a csv? Basically this would affect my workflow - do I keep the data in SAS or do I export an in-database R solution.
11-05-2015 06:28 AM
SAS does not "force R to read a csv." The data is transferred in binary form. I don't know the details or how it compares in speed with the databases you mention, but I know that the developers paid attention to performance.
As you are making your decisions, be sure to read about all the data types that SAS automatically converts between SAS and R.. Also read about the 12 advantages to calling R in the SAS/IML language.
11-05-2015 07:51 AM
The tools in the IML that support R will allow you move data between SAS datasets and R data frames.
That makes me believe they are limited in size to the amount of memory that the R process can use.
There are tools from Teradata that allow you to create data frames that keep the data in the database. I am not sure if the SAS interface to R would work with that implementation of R, but I don't see why it wouldn't.