Arcgisbinding package install error in R

[Barty]

Hi all,

I am unable to install the arcgisbinding package in R. I'm running x64 and have tried with multiple versions of R (3.6, 3.6.3, 4.0, and 4.0.2) using the following fix posted on other blogs

install.packages("arcgisbinding", repos="http://r-arcgis.github.io/r-bridge", type="win.binary")

and no matter which version I use I receive the following error: Warning in install.packages :
package ‘arcgisbinding’ is not available (as a binary package for R version XXX).

Any thoughts here? THanks!

[DanPatterson · Accepted solution]

R Arcgisbinding :: Anaconda Cloud from the esri Anaconda site?

[Anonymous User]

Barty, 

Another way you can safely update your arcgisbinding version is using our setup tool in ArcGIS Pro.

Projects --> Options --> Geoprocessing --> R-ArcGIS Support Tab. There you can pick the version of R you want to use and click on the arrow that will either display "Update package from the internet" or "Install package from the internet" depending on whether you have a pre-existing version of arcgisbinding.

If you are updating, I recommend it now because we just released v244, please make sure you do not have any active R instances while making the update.

Orhun

[MadelineBerger]

Hi Orhun,

I just tried the above solution for the most recent version of R (4.2.1) and had the same issue as Barty (error message "package arcgisbinding is not available as a binary package for this version of R". Any chance there is a new update?

Thanks,

Maddie

[AdamR]

I'm getting this same issue with R4.2.2, any resolution here?

[AdamR]

I was able to fix it for us by doing a complete reinstall of R into a separate directory and pulling arcgisbinding v301 (not listed on esri).

The biggest issue with the entire package is it refuses to accept external package repos as directories. We do a base install of R and then users instigate a separate repo for extra packages. The 'install arcgisbinding script' doesn't have any code to check this. In fact, even if you do an entirely separate R install and point ArcGIS at it, it defaults back to C:/Program Files/R

The trick was to edit the .pyt and adjust the rtools variable to point at the external repo. Editing the registry to point at another RInstall also did not work. Finally, the .pyt has general code errors and spelling mistakes, probably needs a good once-over. Maybe set actual variables for users to select R locations/repos rather than try through Rtools.

[AndrewPadilla]

I am also having this issue. Is there a solution other than suggested by @AdamR, by changing the .pyt and pointing to a new instance of R?

[RiceAdam]

Do you have a standard R 4.2.2 deployment, they're usually under "C:/Program Files/4.2.2" or "C:\Users\USER\AppData\Local\R\win-library\4.2"

If you can install packages there, it should work. Sometimes installing the package via Tools - > Geoprocessing doesn't seem to correctly register arcgisbinding. Double-check it's correct by running through the four install scripts, they all need to successfully complete to work.

If the inital install doesn't work/can't find the arcgisbinding (or call out to the internet), you can dump the arcgisbinding.zip into the arcgisbinding script install directory and it'll fallback deploy from there.