An embedded and charset-unspecified text was scrubbed... Name: not available URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20130202/f6377a51/attachment.pl>
Reading .rdata files
2 messages · Andrew DeCarlo, Duncan Murdoch
On 13-02-02 2:23 PM, Andrew DeCarlo wrote:
Hello, I am a developer working on extracting information from .rdata files using java. I have been looking through the documentation a lot trying to figure out how to read these files, but am a bit stuck at the moment. I have generated two .RDATA files from StatTransfer and have noticed that they both begin in similar ways. *File 1* RDA2 A 2 131585 66560 1026 1 4105 10 .Traceback 254 1026 1 4105 *8 <- these are the only difference* SPSSTest <-FileName 787 * * *File 2* RDA2 A 2 131585 66560 1026 1 4105 10 .Traceback 254 1026 1 4105 *9 **<- these are the only difference* statatest <-FileName 787 Because of the remarkable similarity between these two beginning segments I assume that this is some sort of header. After examining the R Internals Manual (Section 1.8) I was able to discover that the first line describes the file RDA2 means its an ascii data file opposed to XDR or word order binary and the 2 indicates that it is version 2 of R. After this first line there are supposed to be 3 integers that also describe the version. However this is pretty much where section 1.8 ends and even though I searched through the rest of the document (and a lot of Google) I couldn't find anything about what the rest of the bytes/numbers indicate. I may just be missing something (I have seen a lot of people reference the Internals Manual) but is there any documentation on the different data files (RDA, RDX, RDB) and how they are organized. Any info would be greatly appreciated. Thanks a bunch!
You could try the source code: the files are written by the code in https://svn.r-project.org/R/trunk/src/main/serialize.c The comments in there describe it, and the code does what it does. Duncan Murdoch