When using datashield.assign, no message is returned if trying to assign variables which don't exist

[timcadman]

How to reproduce

demo_url <- "https://armadillo-demo.molgenis.net/"
demo_token <- armadillo.get_token(demo_url)

builder <- DSI::newDSLoginBuilder()

  builder$append(
    server = "armadillo",
    url = demo_url,
    profile = "xenon",
    driver = "ArmadilloDriver",
    token = demo_token
  )
  
logindata <- builder$build()
conns <- DSI::datashield.login(logins = logindata, assign = F)

datashield.assign.table(conns, "testdata", "xenon-tests/survival/veteran", variables = c("doesnt_exist", "also_not_there"))

Expected behaviour

A warning is returned stating something like:
Warning: variables "doesnt_exist", "also_not_there", do not exist in table "veteran"

Actual behaviour

ds.colnames("testdata")
$armadillo character(0)

[ymarcon]

I think it is the backend (Armadillo) implementation responsibility: there is no way to retrieve the column names at this point, as this function is at a higher level, in dsBase.

[timcadman]

Ah ok, yes I wasn't sure where this sat. Does opal return a more informative error message if variables are not available?

[ymarcon]

No, opal does not consider this to be an issue, because (1) this is more a filter than an explicit assignment and (2) there could be the case where a dataset is missing a variable on a server because it was not possible to harmonize such variable with their population and this is ok.

[timcadman]

Would you consider implementing a warning in opal? As a researcher it would be quite useful - currently you assign variables and then have to check whether what you specified in vars was actually assigned. A warning indicating that some variables weren't available would be nice. I can implement this in Armadillo, but I'd rather not diverge the functionality between opal and armadillo.