csvsummary: show N most common values in columns

Created by: onyxfish

"value: count" (R-style)

Instead of a random sample. This is more useful.