[FEAT] `sample-longest`

[VladimirAlexiev]

csvtk already has a few commands to sample data: sample, corr, summary.
But I want a new one :-) It's useful to understand the "meaning" of cols by looking at "interesting" values in each col.

return the longest value of each column.

Usage:
  csvtk sample-longest [flags]

Flags:
  -n  --number int       print N rows (default 1)
  -N  --max-number int   sample from first N rows (default all)

This is useful to find "interesting" values in each col, in order to understand the "meaning" of each col.
Beware that there will be no correlation between the values in each returned row!

We have a python script for this, if it's of interest to someone.