I want to commit a clusterisation process on a dataset, however the issue is that some variables have only empty values. As the dataset has over 2k variables It is difficult to do remove them by hand so I created another dataset with numbers of missing values in each variable, transposed it, sorted ascending by value and transposed it back again. How do I order the first dataset by the same order of variables as second dataset so that I can remove first n number of variables that are empty?

I have macros such as DROPMISS however they only remove some of the missing variables so I feel like that is the only way to remove all variables with only missing values. Unless there is a way to cluster variables with such variables present.