[R] How to upgrade library from R 1.9.1 to R 2.0.1

Gabor Grothendieck ggrothendieck at myway.com
Thu Feb 17 05:26:32 CET 2005


YiYao_Jiang <YiYao_Jiang <at> smics.com> writes:

: 
: Dear All:
: 
: 	I have a library for R 1.9.1, it is very easy to setup a library in R 
1.9.1.
:     For example:
:     I want to setup a library "test" for R1.9.1.
:     1. Create a folder "test" in the direct X:\ R\ rw1091\ library \.
:     2. Create a file "DESCRIPTION" in the direct X:\ R\ rw1091\ library\ 
test\.
:     3. Modify the "Package:" and "Title:" item to "test" in the 
file "DESCRIPTION" .
:     4.Create a folder "R" in the direct X:\ R\ rw1091\ library\ test\ .
:     5.Create a file "test" in the direct X:\ R\ rw1091\ library\ test\ R\ .
:     6. Add all the function into the "test" file.
:     Then I can call the "test" using command " library (test) ". 
:     I setup the library using the same way in R2.0.1, but the library can't 
work.
:     The error message is " Error in library(test) : 'test' is not a valid 
package -- installed < 2.0.0? "
:     I search the help file and document but still can't upgrade the library. 
:     Who can tell me how to upgrade library from R 1.9.1 to R 2.0.1 or give 
me a sample, thanks.

I assume you are referring to a package that you wrote and want to
upgrade.

I think there may be confusion here between the source and installed
package.  The source package is what you wrote and it should _not_ go
into \R\rw2001\library\test.    Lets say we put that in \Rpkgs\test .
Then, assuming you just installed R 2.0.1, use the Rcmd.exe program
to install the source package into your R tree. 

cd \Rpkgs
\R\rw2001\bin\Rcmd install test

to install it.  Fix up any errors you get during the install
and repeat until it installs. Now go into R and test it out.  
If it seems to work try running 'Rcmd check' on it:

cd \Rpkgs
\R\rw2001\bin\Rcmd check test

and fix up any errors and repeat until it passes Rcmd check and
reinstall.

See the Writing Extensions Manual if the above is not sufficient.




More information about the R-help mailing list