I need an rscript

I work with sparse matrices: 2 to 10K columns, ~15K rows. I need to count unique values in the data portion of each column. The result need to be graphed as ln (log e) unique value vs ln frequency of each unique value. In most analyses there will be multiple unique values with the same frequency. I need to graph (with leastsquares fit) only the left-most (lowest value). I need to repeat the unique value abstraction on the derived "value by value frequency" data and generate the frequency of each of these as well; will need to plot as ln by ln. Am attaching an example in excel. Column A is a column from dataset. Col B is unique values from ColA. Col C is the frequency of each, and D and E are the ln values. Col F is the uniques from Col C. Col G is the frequency of the uniques in data from Col C. Col H is the first occurrence of the column F value in Col C. Col I is the ln of Col H and used fro the simplified data. Col J and K are the ln for Col F and G (used for final graph).