在大数据表中，将多个列的.N子集

I have a large data.table that contains millions of rows and 30 columns. The columns contain a varying number of categorical features. I would like to remove any features that occur less than a certain proportion. that contains million