帧列的子集，以最大化“完整”的观察

I have a data frame with on the order of 20 numeric columns, each containing significant amounts of NA values. I would like to select a subset of these columns that will give me the most rows containing zero NA values. An exhaustive search would take a lot of computing time--is there a better way to get an approximation? I have a data frame with on the order of 20 num