Relaxing R's naming conventions in mlr3

[Fred-Wu]

I am trying an old project using mlr3 for learning purposes. The data has hundreds of variables directly downloaded from a database, so the names of those variables may not necessarily follow R's naming conventions. It is also inpractical to rename them. With many of those variables with non-R names, here is the error message I got.

Error in .__Task__initialize(self = self, private = private, super = super,  : 
  Assertion on 'feature names' failed: Must have names according to R's variable naming conventions, but element 29 does not comply.

Also this error will definitely occur if I tries to re-encode data with sparse matrix, like

dmy <- caret::dummyVars(" ~ .", data = df)
df <- Matrix(predict(dmy, newdata = df), sparse = T)

This will create more unconventional variable names.

Could mlr3 relax the naming conventions?

[be-marc]

Are you using non-ascii characters?

library(mlr3)
library(data.table)

options(mlr3.allow_utf8_names = TRUE)

data = data.table(target = 1, Æ = 1)
as_task_regr(data, target = "target")

Note,

"mlr3.allow_utf8_names": If set to TRUE, checks on the feature names are relaxed, allowing
non-ascii characters in column names. This is an experimental and temporal option to
pave the way for text analysis, and will likely be removed in a future version of the package.

[Fred-Wu]

Some of the longer names have a space or a hyphen in between.
But I think I can remove those from the names.

But it is good to know there exists an option to relax the naming conventions.