Closed waynelapierre closed 3 years ago
Unfortunately we cannot do that. If a utility doesn't have na.rm it means that I wasn't able to find a faster version using na.rm. i am not the expert to answer about statistical functions.
Finally I have found a faster algorithm for na.rm in colmeans and colsums.
colmeans does not have an na.rm argument while colMedian has one. It makes more sense to use it consistently. In addition, to make the Rfast functions applicable to data.frame, I think it is better to have an na.rm argument for all the functions. In my use case, most data have missing values.