Run orca from Microsoft SQL Server R Services

[sailera19]

Because the old static image export with webshot is deprecated, I am trying to change my code to use orca. The code runs on a Microsoft SQL Server 2017 with Machine Learning Services 9.3 installed. R Version is 3.4.3. Issue also occurs when I try to use Python instead. When I execute the code like this it runs forever. `EXECUTE sp_execute_external_script @language = N'R' , @script = N' library(plotly) if (!require("processx")) install.packages("processx")

p <- plot_ly(z = ~volcano) %>% add_surface()

orca(p, "surface-plot.png") ';` I suspect that it has something to do with the exit code behavior of the command line tool: When calling the same code from RTerm on the same machine with the same user it is working.

[jonmmease]

Hi @l0wsk1ll3d, Thanks for reaching out. Unfortunately this is a setup that the orca devs won't be able to reproduce very easily.

@cpsievert, do you have any thoughts on the R component of this scenario?

[sailera19]

If you need any help on reproducing or testing I am happy to assist.

[cpsievert]

Might be a PATH issue? What does Sys.which("orca") give you in that environment?

[sailera19]

EXECUTE sp_execute_external_script @language = N'R' , @script = N' OutputDataSet <- as.data.frame(Sys.which("orca"));'; returns me "C:\PROGRA~1\orca\orca.exe", which is the correct location of orca on this server.

[antoinerg]

@l0wsk1ll3d To further assist you, it would be useful to report the process' standard output and standard error you are getting when making your system call to Orca. It would provide us with more information on the error you're getting.