cluster_by = "performance", and serialize(data, node$con) in Windows

[Borghis]

Project Robyn

Describe issue

For the past 3 months I'm learning Robyn's potential, but I encountered some problems:

1. In robyn_outputs(), when using cluster_by = "performance", returns error:
   "Error in distinct(x, .data$solID, starts_with("nrmse"), .data$decomp.rssd, :
   ℹ In argument: starts_with("nrmse").
   Caused by error:
   ! starts_with() must be used within a selecting function.
   ℹ See ?tidyselect::faq-selection-context for details."
   And proceeds to create the output of each model selected by the pareto front (more than 100 models).

2. The second issue I encountered is about parallelization in Windows 11 (I used Robyn also in Mac and I never had this error): serialize(data, node$con)

If you can give me some solutions, tips or documentation, I would be very grateful. Thank you in advance!

Environment & Robyn version

Robyn version: 3.11.1
R version: 4.3.2

[gufengzhou]

cluster by performance is actually deprecated. for windows it's always very individual and tricky (multi-core vs multi-thread e.g.). hope it's not a blocker for you?

[Borghis]

Hi, thank you for the reply. I searched for other users with my same problem with parallelization in Windows and they suggested that using 1 core (not using the parallelization) fixed it, and it did but it's really slow.
If there will be any updates regarding this issue please let me know. : )
Thank you again!