无法从R.Net引用外部R程序包

Question

To leverage R from my .Net website, i'm using R.Net (1.5.13). To some extend this is working fine for me, but when i'm trying to reference an external R package (the package: tm) i'm running into deep trouble.

The following C# code:

REngine engine = null;
REngine.SetEnvironmentVariables();
engine = REngine.GetINstance();
engine.Initialize();

engine.Evaluate(".libPaths('C:/Users/Laptop/Documents/R/win-library/3.1')");
engine.Evaluate("library(tm)");

is resulting in: package or namespace load failed form 'tm'

From RStudio the command library(tm) works without any exceptions.

I tried the following steps to solve this problem:

• a reinstall of R
• updating packages from R studio
• adding 'C:/Users/Laptop/Documents/R/win-library/3.1' to R_LIBS, R_LIBS_USER and R_LIBS_SITE

But nothing seems to help so far.

Can someone help me with this nasty problem?

Some information from sessionInfo() :

R version: 3.1.0 (2014-04-10)

Platform: x86_64_w64_mingw32/x64 (64-bit)

other attached packages: [1] tm_0.5-10

Answer 1

The last two issues logged at the codeplex site, here and here , relate to this problem: running R.NET from an IIS web app is a known long-standing issue. Thankfully, recently skyguy94 has done the best diagnosis and explanation of the issue to date, and as I understand proposes a workaround. You should give it a try.

Future versions of R.NET will likely offer an easier way to, er, accomodate IIS, but this is subject to the time of already busy volunteers. Meanwhile, I'll update the home page to flag the vexed issues running under IIS.

Answer 2

There might be other base package/s that may be failing after the package shown in the error.

In my case matrixStats was failing but in the background stats was failing to load, and that was failing because it was trying to find core R dlls inside the stats library.

I fixed the issue from another post: http://author-entry17.rssing.com/chan-9456546/all_p43.html

You have to copy over the dlls from the architecture folder into the package path as a work around.

ie copy dlls in C:\\Program Files\\R\\R-3.1.0\\bin\\i386 or C:\\Program Files\\R\\R-3.1.0\\bin\\x64

to

R\\R-3.3.1\\library\\stats\\libs\\x64 or R\\R-3.3.1\\library\\stats\\libs\\i386

as a work around.