R function: library

List of installed packages

Called without an argument returns a list of installed (not necessarily loaded) packages. Compare with installed.packages.

List of installed datasets

library(help="datasets") gets a list of datasets that are installed with the current R installation.

Non standard evalution

library(pkg_name) uses non standard evaluation (NSE) when evaluating the name of the package to to be loaded. This is why pkg_name needs not be quoted with apostrophes or single quotes.

While NSE safes two characters, it is a constant source of confusion when a package is attempted to be loaded whose name is stored in a variable. The following snippet causes the error Error in library(pkg) : there is no package called ‘pkg’. Execution halted

packages_to_load <- c('ggplot2', 'sqldf');

for (pkg in packages_to_load) {
   print(paste('Trying to load', pkg));
   library(pkg);
}

Github repository about-R, path: /functions/library/NSE/does-not-work.R

In order to specify the package to be loaded with a variable, the character.only argument needs to be set to TRUE:

packages_to_load <- c('ggplot2', 'sqldf');

for (pkg in packages_to_load) {
   print(paste('Trying to load', pkg));
   library(pkg, character.only = TRUE);
}

Github repository about-R, path: /functions/library/NSE/character.only.R

It is also possible to load a package from a variable with a combination of eval and bquote:

packages_to_load <- c('ggplot2', 'sqldf');

for (pkg in packages_to_load) {
   print(paste('Trying to load', pkg));
   eval(bquote(library(.(pkg))));
}

#
#  Show loaded packages
#
(.packages());

Github repository about-R, path: /functions/library/NSE/bquote.R

Another possibility is to use do.call and as.name:

packages_to_load <- c('ggplot2', 'sqldf');

for (pkg in packages_to_load) {
   print(paste('Trying to load', pkg));
   do.call('library', list(package = as.name(pkg)));
}

Github repository about-R, path: /functions/library/NSE/as.name.R

R function: library

List of installed packages

List of installed datasets

Non standard evalution

See also