I want to build a package that involves loading data from mysql using different packages depending on the user's system.
For a windows user it would be through an ODBC connection via package RODBC, while a linux/mac user would use the RMySQL package.
In a script, the following works very well:
if(.Platform$OS.type == "unix") {
library(RMySQL)
} else {
library(RODBC)
}
Now I would like to have these packages loaded at the loading of my package.
I would normally add it in the DESCRIPTION file, under 'Depends:', but this doesn't allow the optional clause.
What is the best way to handle this ?
I think the usual way to solve this is via the .onLoad function (see ?.onLoad or help(".onLoad")).
Section 1.6.3 of the Writing R Extension Manuals gives an overview. Perhaps someone else can point you to a good example, I haven't used it so far.
Related
This bounty has ended. Answers to this question are eligible for a +100 reputation bounty. Bounty grace period ends in 4 hours.
antonio wants to draw more attention to this question:
Explain what is the proper workflow when you are building two packages, where one depends on the other. What is the correct way to run devtools::check() on the depending-on package?
I want to add a local package dependency using R devtools.
The suggested way to add packages to the package DESCRIPTION file is one of the two functions use_package() and use_dev_package(), from the usethis package. The latter adds a dependency on an in-development package. The function help shows the following prototype:
use_dev_package(package, type = "Imports", remote = NULL)
where remote is
a character string to specify the remote, e.g. ‘"gitlab::jimhester/covr"’, using any syntax supported by the remotes package.
The remotes vignette shows the following
# Local
Remotes: local::/pkgs/testthat
So the command should be along these lines:
use_dev_package(foopack, type = "Imports", remote = "local::<foopack>")
However, what should be the path to the foopack. An absolute one or relative to project dir? The root package directory or the R directory with the code, or perhaps the foopack.tar.gz build?
All attempts failed for me.
Needless to say that, beyond having the local dependency properly listed in the DESCRIPTION file, I need it to be seen by the devtools build & check functions.
Edit
As regards use_dev_package(), I found a solution:
if I use devtools::check(), then the dependency appears in the search path, and use_dev_package() does not complain any more (see answer below).
However, it is still unclear to me what arguments should I use to make a development check() for the main package, in particular when the package has a vignette.
Ideally, I should be able to pass the check with local dependencies by passing cran = FALSE, but this still gives
"Package required but not available".
It seems that I have to check the local dependencies before adding them to the description file.
devtools::check("path/to/foopack")
usethis::use_dev_package("foopack", remote ="local::path/to/foopack")
The paths can be relative or absolute, and even a single colon works.
It might be worth noting that, when I build the main package, I can use the ordinary:
devtools::build()
but, for a successful check, I need to use the remote argument:
devtools::check(remote = TRUE)
I can't see a rationale for restating what is in the DESCRIPTION file, but I do not have enough expertise to say it's a bug.
Let's see what the others say in this regard.
Edit
Unfortunately, it seems that the remote argument above does not apply to vignettes. So, if I add a vignette to the package, checks fail with local packages
Until an actual solution is found, all I can do is (sadly) to ignore vignette checks:
devtools::check(remote = TRUE, vignettes = FALSE)
I keep getting the following error on app.powerbi.com:
"Error in pivot_wider(data, names_from=names,values_from=values): could not find function "pivot_wider""
I load the 'tidyr' package as a required library.
I get the same error when using "spread" instead of pivot_wider.
When I specify the function using:
tidyr::pivot_wider(data, names_from=names, values_from=values)
I get a different error:
"Error in loadNamespace(name): there is no package called 'tidyr'".
PowerBI Desktop everything works perfectly fine. It's only on the published PowerBI report that I get the error.
What can I do to resolve this issue?
Alternative library / function to 'pivot_wider' or 'spread'?
I was able to verify using print statements that the report on powerbi.com was not able to access the r packages that were installed.
I was using a .libPaths()[3] to point to the location where my libraries were installed, but this didn't work in powerbi.com.
I ended up re-installing R to a location where I could install packages to, since admin rights are necessary to write to the C:/users/program files location.
I am running an R script from Terminal (not RStudio) and at some point I need the user to write the username and password to access an API. For this, I am using a non-safe approach with the following code:
cat('Write username: ')
username<-readLines("stdin", n=1)
cat('Write password: ')
password<-readLines("stdin", n=1)
I have seen there is a nice implementation in RStudio using the rstudioapi package and the askForPassword function but since I am running my script from Termianl, I wonder if there is any alternative or more secure way to ask for this information
getPass is your best solution.
It supports RStudio and fallsback to tcltk and other backends if missing.
password <- getPass::getPass() is your call.
I am trying to setup the working directory-using the url-"https:\..."
Code:
box_setwd(url)
Error:-
Error in box_id(dir_id) :
box.com API ids must be (coercible to) 64-bit integers.
I am exploring bit64 package meanwhile it will be great to have a workaround.
I hope the details are enough.
the dir and other ids pertaining to the boxr commands are the numeric extensions in the url- it's available in the documentation
In RStudio, when I check and uncheck a package, I see the following commands.
library("ggplot2", lib.loc="~/R/win-library/3.4")
detach("package:ggplot2", unload=TRUE)
Can someone explain what is unload=TRUE does?
Conceptually is there a difference between loading/unloading vs attaching/detaching?
From R's official help pages (see also R Packages - Namespaces):
Anything needed for the functioning of the namespace should be handled at load/unload times by the .onLoad and .onUnload hooks.
For example, DLLs can be loaded (unless done by a useDynLib directive in the ‘NAMESPACE’ file) and initialized in .onLoad and unloaded in .onUnload.
Use .onAttach only for actions that are needed only when the package becomes visible to the user (for example a start-up message) or need to be run after the package environment has been created.
attaching and .onAttach
thus means that a package is attached to the user space
aka the global environment
usually this is done via library(pkg)
and you can use normal fun() syntax
loading and .onLoad
thus means that package is (in any way) made available to the current R-session
(e.g. by loading/attaching another package that depends on it or by using pkg::fun() syntax the first time)
though you will not find functions in the global environment
you can use pkg::fun()
detach related to package environment(which more related to user)
unload = TRUE related to namespace environment(which more related to other package)
after detach, you can not use any function inside that package directly
but unloadnamespace won't prevent you from calling that function,but the other packages can't use its function directly