0

I am trying to develop some code using RevoScaleR and noticed that it has options like rowsPerRead and can output xdf files in chunks using the rxImport method. However my use cases I want to apply some custom transformation (to the entire chunk; not just columns) and output the results in fst format instead of xdf (as I find it to be faster).

Is there a way to apply customer transforms of each chunk being read it using RevoScaleR?

xiaodai
  • 14,889
  • 18
  • 76
  • 140

1 Answers1

0

Following the example it's like the below except it doesn't seem to work and just eats up lots of RAM (at least on Windows)

library(RevoScaleR)
a = RxSasData("file.sas7bdat")
rxOpen(a)
rxNext(a)
rxClose(a)
xiaodai
  • 14,889
  • 18
  • 76
  • 140