R Packages Management - gizotso/R GitHub Wiki

Packages

library(help = splines): documentation on package 'splines'

library(stats): load package stats

detach("package:devtools", unload=TRUE) ! unload a package

When you install a package, and both the global environment (that is, your workspace or current R session), and the installed package, have a function with the same name, R will ?mask? the function and use the function in the highest-level environment.

Load a Package

library(tools)
# attach(tools)

# if package is not in the Lib Search path
library("zoo", lib.loc="C:/Tools/R/RLibs")

require(splines) # package
detach(package:splines)

# or also
library(splines)
pkg <- "package:splines"
detach(pkg, character.only = TRUE)

library("stats4") # loads the package  "stats4"
library(h=stats4) # gives help for all functions
data(package=”stats4”) # gives the list of all available datasets

detach("package:pretty") #detache a package from the current environment

R Repositories

Repositories are listed under the config file %R_HOME%/etc/repositories Rgui() : choose repositories (CRAN + CRAN extra by default)

R CRAN : The Comprehensive R Archive Network

This is the main repository for R

CRAN mirror can be selected in Rgui : "Choose Cran Mirror"

  • chooseCRANmirror()
  • chooseCRANmirror(graphics=FALSE)

Or it can be set in cmd line or at startup

# set a CRAN mirror at startup
# https://www.r-bloggers.com/permanently-setting-the-cran-repository/
local({r <- getOption("repos")
       r["CRAN"] <- "https://cloud.r-project.org"
       options(repos=r)
      })

Package Management

This is the most epic stuff in R as package are actively developed and updated by a large community. Main issues

  • Some packages are abandoned and removed from the CRAN repo, thus even if running the old R version that was working fine with the package you will not be able to install it easily (std commands)
  • new R versions : when upgrading to a major R version (R 3.4, 3.5, 3.6, ...) packages should be installed again and need to be compiled against the correct R core (3.4).
  • In some cases we can also need to stick to a specific version because of constraints (production server, cloud service: Power BI service still running 3.4.4 as of 2020-03-14, ...) but when installing a old R version

Installed Packages (R libraries)

# list installed packages
installed.packages()

# listing of all available packages
dir(.libPaths())

# listing attached packages
search()

ip <- installed.packages(fields = c ("Package", "Version"))
ip[, c("Package", "Version")]
ip["graphics", c("Package", "Version")]

ip = installed.packages()[,1-6]
ip =  as.data.frame(installed.packages()[,c(1,3:6)])

# Default : base and recommanded packages
ip_R_default <- ip[!is.na(ip$Priority), 1:3, drop=FALSE]

ip = as.data.frame(installed.packages())
# colnames(ip) to see all available columns
# package, LibPath, Version, Priority, Depends, Imports, ...

ip = as.data.frame(installed.packages()[,c(1,3:4)])
rownames(ip) <- NULL
# packages that comes with R
ip_default <- ip[!is.na(ip$Priority),1:3,drop=FALSE]
# ignoring packages that comes with R (priority = base/recommended/...)
ip <- ip[is.na(ip$Priority),1:2,drop=FALSE]
print(ip, row.names=FALSE)

# open a window with package listing by library (all packages available)
library()
library(lib.loc = .Library)
  # Packages in library 'C:/Tools/R/R-3.2.5/library':
  # ...

Package Version and Description (on installed packages)

packageVersion("tools")
packageDescription("stats")
packageDescription("stats", fields = c("Package", "Version"))
packageDescription("stats", fields = "Version")
packageDescription("stats", fields = "Version", drop = FALSE)

Checking Dependencies

library(tools)
# check dependencies are installed 
package_dependencies('Hmisc')

ip<-installed.packages()
ip["latticeExtra", c("Package", "Version", "Depends", "Imports")]

# deprecated function in devtools
library(devtools) 
package_deps("devtools")

Available Packages and its attributes (version, depends, ...) (CRAN mirror)

packageStatus()

# available packages : package, Version, Priority, Depends, Imports, ...
ap <- as.data.frame(available.packages())

ap["devtools",c("Version","Depends","Imports","Suggests")]

ap["Hmisc",]$Depends

my_package_list=c("devtools", "Hmisc", "reshape2")
my_package_list_depends = ap[my_package_list, c("Version","Depends","Imports","Suggests")]

ap <- available.packages()
ap["ggplot2","Depends"]

# check available packages in list
ap = available.packages()
ap = ap[ap[, "Package"] %in% nip, c(1:5)]

Installing Packages

Note : Depending on install date and order, dependencies installed will not be the same for a given package. Install can be made from Rstudio or from Rgui (menu/cmds). Ih having issues in RStudio, check if it works in RGui.

Install in a specific R library (usefull to keep separate default packages from additionnal and also when not having root access / Run as admin). Default packages are under .Library -> $R_HOME/library

Install packages

The normal method to install packages is to use the install.packages() function. This will either lookup for an existing binary or download an archive (source) and build the package. CRAN only provides binaries for one version of R prior to the current one

install.packages("ggplot2")

# install under a given lib
install.packages("ggplot2", lib=Sys.getenv('R_LIBS'))
library(ggplot2, lib.loc="/data/Rpackages/")

# install from a specific CRAN mirror
install.packages('RMySQL', repos='http://cran.us.r-project.org')
# dependency option
install.packages('knitr', dependencies = TRUE)

install.packages(c("rj", "r.gd"), repos=http://download.walware.de/rj-1.1)
install.packages("rattle", repos="http://rattle.togaware.com", type="source")

# Windows users may also run the following scripts to install required libraries
source("http://www.math.csi.cuny.edu/pmg/installpmg.R")

Install packages from source

install.packages(path_to_file, repos = NULL, type="source")

install.packages("RJSONIO", repos = "http://www.omegahat.org/R", type="source")

install.packages("C:/Tools/R/RLibs-3.2.5/latticist_0.9-44.tar.gz", repos = NULL, type = 'source')
install.packages("D:/PortableApps/R/R-3.5.3-LibsSrc/latticeExtra_0.6-28.tar.gz", repos = NULL, type = 'source')

install.packages("http://cran.r-project.org/src/contrib/Archive/RNetLogo/RNetLogo_0.9-6.tar.gz", repo=NULL, type="source")
install.packages(http://cran.r-project.org/src/contrib/Archive/ggplot2/ggplot2_0.9.1.tar.gz, repos=NULL, type="source")

Behind the scene, install from source File

Package is downloaded in a Temp folder and a R CMD is run to build the package from the source

download.packages()

dir_temp <- tempdir()
dir.create(dir_temp)
pkg <- download.packages("car", dir_temp, type = "source")

wget http://cran.r-project.org/src/contrib/Archive/ggplot2/ggplot2_0.9.1.tar.gz
R CMD INSTALL ggplot2_0.9.1.tar.gz

R CMD INSTALL /usr/local/R-pkg/package_1.0-2.tar.gz
R CMD INSTALL RJSONIO_0.2-3.tar.gz
R CMD INSTALL --build RJSONIO_0.2-3.tar.gz
R CMD INSTALL --library="C:/Tools/R/Lib" gWidgetsRGtk2_0.0-84.tar.gz

Using Binaries

As an alternative, binaries can be downloaded from the CRAN but will probably require latest R Version. Just need to download and unzip the library archive under the R library. :warning: windows binaries are maintained only for most recent R versions (e.g. on 2020-03, we find binaries from R3.2 up to R4.0)

Since March 2016, Windows and macOS binaries of CRAN packages for old versions of R (released more than 5 years ago) are made available from a central CRAN archive server instead of the CRAN mirrors

Custom package install with devtools

require(devtools)
install_version("ggplot2", version = "0.9.1", repos = "http://cran.us.r-project.org")
# https://support.rstudio.com/hc/en-us/articles/219949047-Installing-older-versions-of-packages

library(devtools)
install.packages.zip("https://cran.r-project.org/bin/windows/contrib/r-release/devtools_1.1.zip")

library(devtools)
install_url('https://cran.r-project.org/src/contrib/Archive/latticist/latticist_0.9-44.tar.gz')

install_github("twitter/AnomalyDetection")
install_github("klutometis/roxygen")
install_github("wch/ggplot2")

setwd("C:/Tools/R/RLibs-3.2.5")
install_local("latticist_0.9-44.tar.gz")

Install dependencies with devtools

setwd("C:/Tools/R/RLibs-3.2.5")
library(devtools)
install_deps(pkg = "latticist")

install_deps(".")  
install_deps("/path/to/package", dependencies="logical")

Microsoft Reproducibility

Microsoft R Open offers predictability by default. During installation, the CRAN repository is configured to point to a specific CRAN repository snapshot. For Microsoft R Open 3.5.3, this fixed CRAN repository snapshot was taken Apr 15, 2019

Microsoft R Open

R Version R Release R Open Snapshot CRAN Repo link
3.6.3 20200229 https://cran.microsoft.com/snapshot/2020-04-01/
3.5.3 20190311 201905 2019-04-15 https://cran.microsoft.com/snapshot/2019-04-15/
3.5.2 201812 201904 2019-02-01
3.4.4 20180315 201804 2018-04-01 https://cran.microsoft.com/snapshot/2018-04-01/
3.3.3 201604 201703 2017-03-15 https://cran.microsoft.com/snapshot/2017-03-15/

R 3.6.0 (April, 2019), R 3.5.0 (April, 2018), R 3.4.0 (April, 2017)

install.packages( "devtools", repos = "https://mran.microsoft.com/snapshot/2018-06-01/")

Installer function utility wrappers

Install if package is not installed

install_package <- function(x){
  if (!require(x,character.only = TRUE))
  {
    install.packages(x,dep=TRUE)
    if(!require(x,character.only = TRUE)) stop("Package not found")
  }
}

Is installed

p %in% installed.packages()

Using require()

p = "Hmisc"
if (!require(p, character.only = TRUE)) {print(paste(p, "needs install"))}

if (!require(p, character.only=T, quietly=T)) {
    install.packages(p)
    library(p, character.only=T)
}

Upgrade R Packages (not recommended)

update.packages() : updates all the packages installed with the newest version available in the repositories

update.packages(c("ggplot2", "reshape2", "dplyr"))`
update.packages(checkBuilt=TRUE, ask=FALSE)

Remove a Package

remove.packages() : remove a package uninstall.packages() : uninstall

remove.packages()
remove.packages("survival", lib=Sys.getenv('R_LIBS'))

uninstall.packages(c("reshape", "plyr"))

Create your own packages

Upgrade R

InstallR package https://cran.r-project.org/web/packages/installr/index.html

# installing/loading the package:
if(!require(installr)) {
  install.packages("installr")
  require(installr)
} #load / install+load installr
updateR() # updates R

install.packages("installr")
library(installr)
updateR()

Then upgrade moved packages

update.packages(checkBuilt=TRUE)

My Packages

my_package_list=c("devtools", "Hmisc", "reshape2", "dplyr", "gplots", "ggplot2", "ggExtra", "playwith", "gridExtra", "prettyR", "psych", "lessR", "tables", "xtable", "xlsx", "corrplot", "psy", "Epi", "MASS", "foreign", "XLConnect", "RCurl", "htmlTable", "knitr", "rmarkdown", "shiny", "shinyjs", "twitteR", "wordcloud")

## devtools :dependencies stringi, magrittr, BH, mime, R6, stringr, brew, xml2, httr, curl, memoise, whisker, evaluate, digest, rstudioapi, jsonlite, roxygen2, rversions, git2r
## Hmisc : dependencies Formula, latticeExtra, acepack, gridExtra
## latticist : not anymore in R 3.2
## playwith : ## dependency cairoDevice, gWidgetsRGtk2, gWidgets, gridBase
## lessR : ## also installing the dependencies minqa, nloptr, RcppEigen, lme4, SparseM, MatrixModels, pbkrtest, quantreg, car, leaps, MBESS, sas7bdat, readxl, triangle
## rattle : ## RGtk2
## htmlTable : ## dependencies formatR, highr, markdown, yaml, knitr
## shiny : ## crayon, XML, httpuv, htmltools, Cairo, testthat
## Rmcdr :## dependencies zoo, sandwich, e1071, RcmdrMisc, tcltk2, abind