search for: lubinski

Where precisely is the bug here? If you assign a list (d[,1:2] is a list: it is the whole data frame) to a numeric matrix you coerce the latter to mode list, and that is what happened. If we have the simpler version d <- data.frame(v1=1:2,v2=3:4) x <- matrix(0, 2, 2) x[] <- d it may be easier to see what happened. I think the problem is in your subject line: d[,1:2] is _not_ a