[R] exporting list of installed packages for import on another system?

[Marc Schwartz]

I'll leave that up to you...  :-)

You could add the URL of the thread to the wiki entry if you wish as  
well.

Regards,

Marc

On Jul 14, 2009, at 10:57 AM, Jan Theodore Galkowski wrote:

> Super Marc! Thanks!
>
> Should I post this on the R Wiki some place? 'Twould be useful to  
> others, I think.
>
> - Jan
>
> On Tue, 14 Jul 2009 11:49:49 -0400, Marc Schwartz <marc_schwartz at me.com 
> > wrote:
>
>> On Jul 14, 2009, at 9:43 AM, Jan Theodore Galkowski wrote:
>>
>>> Is it possible to export a list of installed packages from WinXP,  
>>> and
>>> use that export to import the same set of packages on Ubuntu  
>>> (Jaunty)?
>>> No doubt
>>> there is custom code that could be written, but I wonder if R  
>>> 2.9.1 has
>>> anything built it to do that?  Is it as simple as moving something  
>>> like
>>> Rprofile.site from one machine to the other?
>>>
>>> I had a look at R-admin.pdf, and although it talks a lot about
>>> configuring on various systems, it did not address this directly.   
>>> Also
>>> looked at RSeek.
>>>
>>> Thanks.
>>
>>
>> If you are just going to replicate a standard installation with  
>> Base and Recommended packages, then just install R on Ubuntu (I  
>> presume that you will use 'apt-get'"?) and you will have the same.  
>> Review the following for more Ubuntu specific information:
>>
>>   http://cran.r-project.org/bin/linux/ubuntu/
>>
>>
>> If there are extra packages that you have installed on Windows,  
>> then you can use the following to get the list:
>>
>> IP <- as.data.frame(installed.packages())
>>
>> MyPkgs <- subset(IP, !Priority %in% c("base", "recommended"),  
>> select = c(Package, Bundle))
>>
>>
>> MyPkgs will now contain a list (first column) of the packages that  
>> you have installed that are not part of the basic R install. In  
>> addition, pay attention to the 'Bundle' column in case you have  
>> installed any package bundles. Those would need to be installed  
>> using the Bundle name and not the individual package name.
>>
>> Before you go too far with this however, I would check to see just  
>> how many packages are listed in MyPkgs. If the list is short (for  
>> some value of short), you may be better just manually installing  
>> the packages on your Ubuntu system rather than going through this  
>> process.
>>
>> The question then becomes, are you going to install these on Ubuntu  
>> using 'apt-get' from the Ubuntu CRAN repos, or are you going to  
>> install the packages from CRAN using install.packages(). I suppose  
>> intertwined with that will be are there any packages that you have  
>> installed that are not yet in the Ubuntu repos.
>>
>> In either case, you can save 'MyPkgs' to an R readable object file  
>> on Windows by using:
>>
>>   save(MyPkgs, "MyPkgs.Rdata")
>>
>> Copy that file over to your Ubuntu installation and use:
>>
>>   load("MyPkgs.Rdata")
>>
>> and you will have the MyPkgs object available there.
>>
>> You can then use the list as you require.
>>
>> If you are going to use install.packages() and presuming that you  
>> do not have any bundles installed on your Windows system, you could  
>> do the following after using 'sudo R' to go into R:
>>
>>   load("MyPkgs.Rdata")
>>
>>   install.packages(MyPkgs$Package, dependencies = TRUE)
>>
>>
>> If you are going to use 'apt-get', I would read the following as I  
>> noted above:
>>
>>   http://cran.r-project.org/bin/linux/ubuntu/
>>
>>
>> You could feasibly create an 'apt-get' command line call using  
>> paste() and the system() functions along the lines of:
>>
>>   CMD <- sapply(MyPkgs$Package, function(x) paste("r-cran-", x, sep  
>> = ""))
>>   CMD <- paste(CMD, collapse = " ")
>>   CMD <- paste("apt-get", CMD)
>>
>> and then use:
>>
>>   system(CMD)
>>
>> after using 'sudo R' to get into R.
>>
>> However, I would recommend that you consider posting a query to the  
>> r-sig-debian list just to verify all of the above. More info at:
>>
>>   https://stat.ethz.ch/mailman/listinfo/r-sig-debian
>>
>> HTH,
>>
>> Marc Schwartz
>>
>
>
>
> -- 
>
>  Jan Theodore Galkowski
>  bayesianlogic at acm.org
>  http://www.ekzept.net
>  16072391834
>
> "Eppur si muove." --Galilei
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.