Love it or hate it, but Microsoft is fairly ubiquitous in any business environment. That could be reliance on Excel spreadsheets and Word documents up to full-blown Azure environments, PowerBI reports, etc.
I have learned that there are some times when R and OneDrive butt heads, so here are some tips for making them get along:
Tip 1: Don’t install R to a OneDrive folder
Newer versions of R don’t suggest this anymore, but be sure not to install R into a OneDrive synced folder like Documents. It sounds like a good idea for portability, but updates will throw a great big wrench into things in a hurry, trust me. This also applies to related software like quarto, rtools, python/conda, really anything that can be installed locally as a normal user.
Tip 2: No R Projects (with some caveats)
Unfortunately, this will also throw a wrench into things. R Studio makes a lot tiny folders with very deep folder trees and you can very quickly run into Window’s character limits. This is a major issue for RMarkdown/Quarto documents, and will break the ability to view code output without rendering. And then when you do render, all of those tiny deeply-nested files get changed and OneDrive loses it’s mind trying to track the changes. In fact, even when not using R Projects, it’s a good idea to wait a little while between document renders for this very reason.
The only real way to use R Projects is to keep them outside of OneDrive folders. That works well, but then you lose the backup features of OneDrive. It really depends on your company’s storage environment.
In lieu of R Projects, the best suggestion I have is to make a new folder for any project you’re working on and keep all scripts and outputs there. It’s not the same as a project but it will get the job done.
Tip 3: Always keep on this device
This option (via right-click in Windows Explorer) is a life-saver for reading data from shared locations. There’s just not a good way that I know of to access documents in the OneDrive/SharePoint cloud, so keeping them synced on your local PC is the only way. reaxl::read_xlsx attempting access a file seems to trigger Windows to download said file, but not always. And if a file is unchanged or unopened (at least as far as Windows knows) for a certain amount of time, that file may get removed locally to save storage space and then your scripts to read data will stop working.