Hi!
Is it possible or would it be possible in future to use the RStudio package manager with the MRAN repos from Microsoft to install packages depending on a specific date?
Bests
shearer
Hi!
Is it possible or would it be possible in future to use the RStudio package manager with the MRAN repos from Microsoft to install packages depending on a specific date?
Bests
shearer
Great question!
RStudio Package Manager uses a dedicated RStudio service as the upstream source for CRAN packages. This service is not the same as a public CRAN mirror (like MRAN) and adds some extra metadata to help RStudio Package Manager track all the changes that occur on CRAN everyday. So the short answer is no, you can not point RStudio Package Manager at an MRAN snapshot.
However, in many discussions with R users and admins, most people's goal in using MRAN snapshots is to ensure reproducibility and help users in the company access a consistent set of packages (so that everyone is using the same package versions). RStudio Package Manager provides a few facilities to meet these goals:
See an example as well as what it looks like to add or update packages.
You probably weren't looking for such a long answer, but let me know if we can provide more information!
You may also be curious in how RStudio Package Manager integrates with different change control strategies.
Hi slopp,
thank you for your fast reply and the interesting answer.
I took a look at your referenced package manager example page and find it very impressive.
If I understand the functionality right I can choose for every package one of the archived versions shown at the bootom of the page.
Since the versions with corresponding release dates are given, I question me, if it would be possible to include a method in a future release which install chosen packages at a given date?
Something like
devtools::install_date('c(A3', 'dplyr', 'rms'), '2017-06-12')
instead of
devtools::install_version('A3', '0.9.1') and so on for the other packages.
I think it would be a nice alternative to MRAN with more flexibility due to the other options.
With
library(versions)
it would be possible to extract versions and release dates from given packages and therewith it would be possible to deternine the version at a given date followed by an install_version() command with this information.
str(available.versions("dplyr"))
List of 1
dplyr:'data.frame': 20 obs. of 3 variables: .. version : chr [1:20] "0.7.6" "0.7.4" "0.7.3" "0.7.2" ...
.. date : chr [1:20] "2018-06-29" "2017-09-28" "2017-09-09" "2017-07-20" ... .. available: logi [1:20] TRUE TRUE TRUE TRUE TRUE TRUE ...
Bests
shearer