R-window Stops Workflow

[sabe]

Dear Alteryx-community members,

I created a workflow including the R-tool and installed several packages (see list below) to the library on C:/ProgramData/Alteryx/R-library. To avoid version conflicts, I load all the packages pre-installed from Alteryx via the default path (C:/Program Files/Alteryx/R-3.6.3/library) first and then the packages from the path mentioned above.

However, when I now open the workflow in Alteryx (Version: 2020.4.5.12471) it takes about 3 min till Alteryx responds and an R-window automatically pops up, which can not be closed. Crucially, to this point, I haven't done anything except opening the workflow.

When I click on "Run", the workflow runs until the same thing happens: The same R-window pops up. Closing the R-window causes the workflow to stop with "Error: createReports: The R.exe exit code (3489660927) indicated an error".

Has anyone experienced this or a similar issue and knows how to fix it?
Thank you very much for your help!


Here is the list of R-packages I try to install and load from C:/ProgramData/Alteryx/R-library:
pdftools, dplyr, tidyr, stringr, purrr, readxl, ggplot2, RCurl, htmltools, pagedown, later, websocket, qpdf, rlang, servr.

[apathetichell]

@mbarone"The bottom line is if you want Alteryx to run R code, via an R tool on the Canvas, then any packages must be installed to C:\Program Files\Alteryx\R-[most recent version]\library." - Have you tried setting a different path for packages using .mypaths?

[mbarone]

No I haven't. Not sure if that would work.  From my understanding, if you reference a package in the tool (or script) from the R tool on the Canvas, it will only look in that default path.

[apathetichell]

sorry - tried to edit the last message to include this - packages installed to and accessible by Alteryx R at multiple locations....

[mbarone]

No, I haven't done that.  Not sure if it'll cause errors or not.  The R-Alteryx "integration" is very fragile LOL

[apathetichell]

@mbaronetotally - it tends to work for me (and allows me to control package install in Rstudio) - still less fragile than the venv of Python...

[sabe]

Thank you for this great explanation!
@apathetichell to answer your question: Yes, I checked the versions and installed only (older) package versions that are conform with R 3.6.3