Skip to content

Windows Installation Without Third-Party Packages

9 messages · Elliot Joel Bernstein, John McKown, Jeff Newmiller +4 more

#
I am trying to install R for Windows, but when I use the installer provided
on CRAN, a number of third-party packages are installed by default (i.e.
lattice, Matrix, codetools, etc.). If R is installed with administrator
privileges, so it's available for all users, non-administrators can't
update those packages. Is there any way to just install R without any
third-party packages, and let individual users install the packages they
want?

Thanks.

- Elliot
#
On Thu, Apr 9, 2015 at 8:42 AM, Elliot Joel Bernstein <ejb6 at cornell.edu>
wrote:
?Please try to not post in HTML, per forum standards.

I don't know if this will help, but I hope so. I think what I did will be
self explanatory
trying URL 'http://cran.rstudio.com/bin/windows/contrib/3.1/plyr_1.8.1.zip'
Content type 'application/zip' length 1154715 bytes (1.1 Mb)
opened URL
downloaded 1.1 Mb

package ?plyr? successfully unpacked and MD5 sums checked

The downloaded binary packages are in
C:\Users\john.mckown\AppData\Local\Temp\RtmpWIjtdm\downloaded_packages
[1] "C:/Users/john.mckown/Documents/R/win-library/3.1"
[2] "C:/Program Files/R/R-3.1.1/library"
[1] "C:/Users/john.mckown/Documents/R/win-library/3.1"
[2] "C:/Program Files/R/R-3.1.1/library"
[1] "C:/Users/john.mckown/Documents/R/win-library/3.1"

...
[116] "C:/Users/john.mckown/Documents/R/win-library/3.1/plyr"

[117] "C:/Users/john.mckown/Documents/R/win-library/3.1/plyr/data"

[118] "C:/Users/john.mckown/Documents/R/win-library/3.1/plyr/help"

[119] "C:/Users/john.mckown/Documents/R/win-library/3.1/plyr/html"

[120] "C:/Users/john.mckown/Documents/R/win-library/3.1/plyr/libs"

[121] "C:/Users/john.mckown/Documents/R/win-library/3.1/plyr/libs/i386"

[122] "C:/Users/john.mckown/Documents/R/win-library/3.1/plyr/libs/x64"

[123] "C:/Users/john.mckown/Documents/R/win-library/3.1/plyr/Meta"

[124] "C:/Users/john.mckown/Documents/R/win-library/3.1/plyr/R"

[125] "C:/Users/john.mckown/Documents/R/win-library/3.1/plyr/tests"

[126] "C:/Users/john.mckown/Documents/R/win-library/3.1/R6"

...

As you can see, "plyr" got installed in my personal area. And it is still
on the system directory:
[1] "C:/Program Files/R/R-3.1.1/library/parallel/tests"
 [2] "C:/Program Files/R/R-3.1.1/library/plyr"
 [3] "C:/Program Files/R/R-3.1.1/library/plyr/data"
 [4] "C:/Program Files/R/R-3.1.1/library/plyr/help"
 [5] "C:/Program Files/R/R-3.1.1/library/plyr/html"
 [6] "C:/Program Files/R/R-3.1.1/library/plyr/libs"
 [7] "C:/Program Files/R/R-3.1.1/library/plyr/libs/i386"
 [8] "C:/Program Files/R/R-3.1.1/library/plyr/libs/x64"
 [9] "C:/Program Files/R/R-3.1.1/library/plyr/Meta"
[10] "C:/Program Files/R/R-3.1.1/library/plyr/R"
[11] "C:/Program Files/R/R-3.1.1/library/plyr/tests"
[12] "C:/Program Files/R/R-3.1.1/library/proto"
Pardon the weird subscript, but the list was way to big to cut, paste, and
edit. So your users should be able to force any package, even a "system"
package, into their personal R directory using the "lib=" parameter of the
install.packages() function. This will allow them to update their copy of
any R package from CRAN.

I hope.?
#
John's answer is correct but you might like one of the following summaries better.

The short (baby bear) answer is no. A longer (papa bear) answer is that anything is possible if you dig deep enough. But the just-right (mama bear) answer is that you don't need to worry about it since users should normally be updating their personal libraries which will take precedence over the system-wide library.
---------------------------------------------------------------------------
Jeff Newmiller                        The     .....       .....  Go Live...
DCN:<jdnewmil at dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
                                      Live:   OO#.. Dead: OO#..  Playing
Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
/Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k
--------------------------------------------------------------------------- 
Sent from my phone. Please excuse my brevity.
On April 9, 2015 11:13:45 AM PDT, John McKown <john.archie.mckown at gmail.com> wrote:
#
My understanding is that the packages installed with the windows installer
were only updated by installing a new version of R or the  patched install
file for the current version. If this is the case you you do not need to be
concerned about updates to these packages. Perhaps some one wiser that I
can confirm if my assumption is right or wrong.

John

John C Frain
3 Aranleigh Park
Rathfarnham
Dublin 14
Ireland
www.tcd.ie/Economics/staff/frainj/home.html
mailto:frainj at tcd.ie
mailto:frainj at gmail.com
On 9 April 2015 at 14:42, Elliot Joel Bernstein <ejb6 at cornell.edu> wrote:

            

  
  
#
On 09.04.2015 23:16, John C Frain wrote:
No, updates of such recommended packages may be available between R 
releases. But individual users can install new versions using 
install.packages() into their private libraries, e.g. those that are 
first on the search path (given by .libPaths()) so that they are loaded 
first.

Best,
Uwe Ligges
#
However, little care is needed since the version in a private library may override the system one even after an R upgrade has updated it; i.e. an older version can end up in front of a newer one, which may cause some confusion.

  
    
#
I would suggest that the scenario suggested by Peter is exactly as it should be... with the user fully in charge of which packages they use. A simple update can correct any version "inversion" if it occurs, and the user need not blame the sysadmin if their scripts stop working because of a system update.
---------------------------------------------------------------------------
Jeff Newmiller                        The     .....       .....  Go Live...
DCN:<jdnewmil at dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
                                      Live:   OO#.. Dead: OO#..  Playing
Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
/Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k
--------------------------------------------------------------------------- 
Sent from my phone. Please excuse my brevity.
On April 9, 2015 2:53:29 PM PDT, peter dalgaard <pdalgd at gmail.com> wrote:
#
Dear Eliot,

Users cannot update those packages because they don't have the correct
privileges to the library directory inside the R installation. If the
administrator gives enough priviliges to the users so that they can modify
all within the library directory, then the users should be able to update
those packages.

Best regards,

ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Research Institute for Nature and
Forest
team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance
Kliniekstraat 25
1070 Anderlecht
Belgium

To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey

2015-04-09 15:42 GMT+02:00 Elliot Joel Bernstein <ejb6 at cornell.edu>:

  
  
#
On 09 Apr 2015, at 23:53 , peter dalgaard <pdalgd at gmail.com> wrote:

            
Aaaargh. 

 _a_ little care is needed.

(The other thing means that you almost do not need to care.)