Hi, I want R to be installed on a UNIX network (Solaris). I am not the system administrator and so I cannot do it myself. The system administrator wants to know which packages I want, and it may be a lot. Is there an easy way to download and install all packages at once? Is it a good idea? (There is a good chance that I do not need the some few packages that do not install well during such a procedure, and so I would like to tell the sysadmin to do it even if it will not be 100% successful). Unfortunately it may be even more compicated, because it may be (I was not able to find it out absolutely surely) that our net architecture does not allow to do it via install.packages from within R. So what is the easiest way to do it from outside R? If it is not a good idea to install all packages at once, what is the easiest way to download and install a list of, say, 30 packages? R CMD INSTALL accepts a list as input; but how to download a list of 30 packages at once? (The sysadmin would really appreciate if we could do it in a way that later additions of packages are reduced to a minimum.) Best, Christian *********************************************************************** Christian Hennig Fachbereich Mathematik-SPST/ZMS, Universitaet Hamburg hennig at math.uni-hamburg.de, http://www.math.uni-hamburg.de/home/hennig/ ####################################################################### ich empfehle www.boag-online.de
I once copied all the contrib package sources from CRAN to a local directory (I used an ftp client, but you could use rsync or wget : just don't do it recursively or you will get the Archive too) and ran R CMD INSTALL *.tar.gz The only overhead in installing all packages at once is disc space: we seem to have 320Mb in R_HOME/library here. On Tue, 7 Oct 2003, Christian Hennig wrote:> Hi, > > I want R to be installed on a UNIX network (Solaris). I am not the system > administrator and so I cannot do it myself. > The system administrator wants to know which packages I want, and it may be > a lot. > Is there an easy way to download and install all packages at once? > Is it a good idea? (There is a good chance that I do not need the some > few packages that do not install well during such a procedure, and so I would > like to tell the sysadmin to do it even if it will not be 100% successful). > > Unfortunately it may be even more compicated, because it may be (I was not > able to find it out absolutely surely) that our net > architecture does not allow to do it via > install.packages from within R. So what is the easiest way to > do it from outside R? > > If it is not a good idea to install all packages > at once, what is the easiest way to > download and install a list of, say, 30 packages? R CMD INSTALL accepts a > list as input; but how to download a list of 30 packages at once? > > (The sysadmin would really appreciate if we could do it in a way that > later additions of packages are reduced to a minimum.) > > Best, > Christian > > > *********************************************************************** > Christian Hennig > Fachbereich Mathematik-SPST/ZMS, Universitaet Hamburg > hennig at math.uni-hamburg.de, http://www.math.uni-hamburg.de/home/hennig/ > ####################################################################### > ich empfehle www.boag-online.de > > ______________________________________________ > R-help at stat.math.ethz.ch mailing list > https://www.stat.math.ethz.ch/mailman/listinfo/r-help > >-- Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595
Christian Hennig wrote:> Hi, > > I want R to be installed on a UNIX network (Solaris). I am not the system > administrator and so I cannot do it myself. > The system administrator wants to know which packages I want, and it may be > a lot. > Is there an easy way to download and install all packages at once?Yes. Combination of CRAN.packages() and install.packages() will do, but it's for sure an overhead to install all packages given you are the only one who uses R on that machine.> Is it a good idea? (There is a good chance that I do not need the some > few packages that do not install well during such a procedure, and so I would > like to tell the sysadmin to do it even if it will not be 100% successful). > > Unfortunately it may be even more compicated, because it may be (I was not > able to find it out absolutely surely) that our net > architecture does not allow to do it via > install.packages from within R. So what is the easiest way to > do it from outside R? > > If it is not a good idea to install all packages > at once, what is the easiest way to > download and install a list of, say, 30 packages? R CMD INSTALL accepts a > list as input; but how to download a list of 30 packages at once?E.g. download.file() within R or "wget" outside are the tools I'd choose.> (The sysadmin would really appreciate if we could do it in a way that > later additions of packages are reduced to a minimum.)Have you told the sysadmin that some packages are updated very frequently? I propose to set up your own library section (in a directory accessible by yourself), where you can install packages independently from the sysadmin. Uwe Ligges
Christian Hennig <fm3a004 at math.uni-hamburg.de> writes:> I want R to be installed on a UNIX network (Solaris). I am not the system > administrator and so I cannot do it myself. > The system administrator wants to know which packages I want, and it may be > a lot. > Is there an easy way to download and install all packages at once? > Is it a good idea? (There is a good chance that I do not need the some > few packages that do not install well during such a procedure, and so I would > like to tell the sysadmin to do it even if it will not be 100% successful).Others have commented on how to do it outside of R. If you end up being able to do it within R, consider the following hack'd function: installNewCRANPackages <- function() { test2 <- packageStatus()$avail["Status"] install.packages(row.names(test2)[which(test2$Status=="not installed")]) } best, -tony -- rossini at u.washington.edu http://www.analytics.washington.edu/ Biomedical and Health Informatics University of Washington Biostatistics, SCHARP/HVTN Fred Hutchinson Cancer Research Center UW (Tu/Th/F): 206-616-7630 FAX=206-543-3461 | Voicemail is unreliable FHCRC (M/W): 206-667-7025 FAX=206-667-4812 | use Email CONFIDENTIALITY NOTICE: This e-mail message and any attachme...{{dropped}}
If it's really true that you will be the only user of R, then it might be easier to build it in some place where you do have the necessary permissions. You would then follow the suggestion in the provided INSTALL file (from R 1.7.1): ---- quote ---- INSTALLATION You do not need to install R to run it: you can run R by the script `bin/R' which you can link or copy to any convenient place in your path. ---- end quote ---- Considering that R itself is updated fairly often, your sysadmin may appreciate if you are able to take care of maintaining R yourself. If you are not the only user, your sysadmin might be willing to make the necessary copy or link in some directory in the standard user's path, leaving the installation in some place you own. Then you can do all the updates and package installations, and the sysadmin only needs to update the links/copies when R itself is updated. That's how we do it here, and it works very well. -Don At 10:27 AM +0200 10/7/03, Christian Hennig wrote:>Hi, > >I want R to be installed on a UNIX network (Solaris). I am not the system >administrator and so I cannot do it myself. >The system administrator wants to know which packages I want, and it may be >a lot. >Is there an easy way to download and install all packages at once? >Is it a good idea? (There is a good chance that I do not need the some >few packages that do not install well during such a procedure, and so I would >like to tell the sysadmin to do it even if it will not be 100% successful). > >Unfortunately it may be even more compicated, because it may be (I was not >able to find it out absolutely surely) that our net >architecture does not allow to do it via >install.packages from within R. So what is the easiest way to >do it from outside R? > >If it is not a good idea to install all packages >at once, what is the easiest way to >download and install a list of, say, 30 packages? R CMD INSTALL accepts a >list as input; but how to download a list of 30 packages at once? > >(The sysadmin would really appreciate if we could do it in a way that >later additions of packages are reduced to a minimum.) > >Best, >Christian > > >*********************************************************************** >Christian Hennig >Fachbereich Mathematik-SPST/ZMS, Universitaet Hamburg >hennig at math.uni-hamburg.de, http://www.math.uni-hamburg.de/home/hennig/ >####################################################################### >ich empfehle www.boag-online.de > >______________________________________________ >R-help at stat.math.ethz.ch mailing list >https://www.stat.math.ethz.ch/mailman/listinfo/r-help-- -------------------------------------- Don MacQueen Environmental Protection Department Lawrence Livermore National Laboratory Livermore, CA, USA