Recently I have been unable to get the Rserve addon to run properly. I am using Windows 7 (64 bit) and both J64-602 and jqt64. I run certain statistical routines in R called through J, by means of a cover function that accepts a list of boxed strings, one for each J command to be executed, runs all but the last with rdcmd, and runs the last, inside the R function capture.output(), with rdget. In this way I am able to get the results I need (e.g. from a GLM) as they would appear in the R session.
Everything worked fine about two weeks ago. Since that time I have updated all my downloaded packages, without paying attention to whether, e.g. stats/r/rserve actually changed. Now my calls to R to run GLM's fail when running the R function summary() - although they run the other commands, pointing R to the data file and actually fitting the model, just fine. What is puzzling is that if I feed R a very simple set of commands, such as 'rn<-rnorm(8)' followed by summary(rn), the summary comes back fine. If I feed R my GLM but instead of summary() ask for residuals(), these come back fine even though the output is large, much larger than the output of summary(). When an R run fails, I trap the context and print out the entire set of commands so I can copy and paste them into an R session. Invariably R runs these commands with no problem, printing the results of summary() to screen. Unfortunately, I would like the results to come back into J rather than having to run them manually in R, since I parse these results and process them further in J. What could have caused this to stop working just in the last two weeks? Any insight would be most appreciated. ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
