warning of rankMatrix() in df2lambda()

[sbrockhaus]

For the computation of df2lambda the rank of the design matrix is computed, see
https://github.com/hofnerb/mboost/blob/master/pkg/mboostPatch/R/bl.R#L9
and
https://github.com/hofnerb/mboost/blob/master/pkg/mboostPatch/R/bkronecker.R#L57

In the case that 'p>n', one gets a warning that the rank is computed for t(x) instead of x:

Warning in structure(if (useGrad) which.min(diff1) else if (Meth == "qr") { :
  rankMatrix(x, method='qr'): computing t(x) as nrow(x) < ncol(x)

I think it would be better to suppress the warning, by setting

rank_X <- rankMatrix(X, method = 'qr', warn.t = FALSE)

or by checking whether nrow(X) < ncol(X) and then supplying, t(X) instead of X, see the help of ?rankMatrix argument warn.t.
(As the warning occurs for each base-learner with 'p<n' it can get kind of annoying for models with many base-learners, especially when using cvrisk() on the model.)

In the following a small example for the warning:

library(mboost)
### setting with 'p>n' gives warning in rankMatrix() 
set.seed(290875)
n <- 10
x1 <- rnorm(n)
y <- 2*x1 
mod2 <- mboost(y ~ bbs(x1, df = 4))
dim(extract(mod2, "design")[[1]])
# options(warn = 1)

### for comparison: setting with 'p<n' does not give this warning
set.seed(290875)
n <- 30
x1 <- rnorm(n)
y <- 2*x1 
mod2 <- mboost(y ~ bbs(x1, df = 4))
dim(extract(mod2, "design")[[1]])

[sbrockhaus]

Thanks! There is another code-line giving the same warning as rankMatrix() is used:
https://github.com/hofnerb/mboost/blob/master/pkg/mboostPatch/R/bkronecker.R#L57
Can you set the argument warn.t = FALSE here as well?
Best