QA: Does RevoScaleR only work with .xdf files, or can data frames be used for analysis too?

Data frames can be used in many RevoScaleR functions. Smaller data sets or excerpts from big data stored in an .xdf file can be read into an R data frame, and all of R's functionality is available for data frames 

This example reads the first 10 rows from the CensusWorkers .xdf file into a data frame: 

censusWorkers<-file.path(rxGetOption("sampleDataDir"),"CensusWorkers") 
myCensusDF <- rxReadXdf(censusWorkers,numRows = 10)

and this example compares rxLinMod and lm results using the iris data: 

irisLinMod <- rxLinMod(Sepal.Length ~ Sepal.Width, data = iris) 
summary(irisLinMod) 
irisLM <- lm(Sepal.Length ~ Sepal.Width, data = iris, contrasts=list(Species = contr.SAS)) 
summary(irisLM)