Hi, Our company is doing processing and analysis of (mostly) quantitative data. Some of our researchers are R users, some are Python users. Our data is accessed via thrift-based services. Currently, we have developed parallel R/Python clients that access the data. This has been done using thrift generated c++ objects in R (accessed via a dyn.loaded shared library) and thrift generated python objects in python. We would like to merge those duplicate threads of development, and have a single client speaking (via thrift) to the data service. There is a fair amount of work marshaling the thrift objects into both python and R and we would like to halve that effort by only doing it once and (somehow) sharing the result across the two languages.
So, I am seeking advice on how to achieve this, possibly using rpy. It is easier working with thrift from python and we want a solid python thrift client for other purposes. So ideally, we would like someway to get the python objects into an R process. Our objects are relatively simple, matrices with accompanying scalar metadata. I have not done a thorough reading of the rpy documentation, but on the face of it the it appears the R engine is wrapped by the python one. My naive approach to bring the data into a running R process would be to first bring the objects into a python process, inject those into R-space using rpy, save that, and then attach to it using R. Transfer performance is not critical, and the data path is one way (server -> client). If anyone has comments or suggestions, that would be greatly appreciated. Matt ------------------------------------------------------------------------------ Nokia and AT&T present the 2010 Calling All Innovators-North America contest Create new apps & games for the Nokia N8 for consumers in U.S. and Canada $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store http://p.sf.net/sfu/nokia-dev2dev _______________________________________________ rpy-list mailing list rpy-list@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rpy-list
