docs(python): improve n_unique and approx_n_unique docs

[alexander-beedie]

Added a note about null being considered a distinct value in its own right for the purposes of n_unique, and extended the examples a little to better illustrate the functionality.

[stinodego]

Good to have this clarified! It has surprised me in the past.

EDIT: Looks the approx_n_unique outcome is not stable - might want to ignore the result of that example.

[alexander-beedie]

EDIT: Looks the approx_n_unique outcome is not stable - might want to ignore the result of that example.

Done! Added a 'skip' directive 👍

[stinodego]

EDIT: Looks the approx_n_unique outcome is not stable - might want to ignore the result of that example.

Done! Added a 'skip' directive 👍

Actually the doctest: +IGNORE_RESULT directive would be more fitting for this (this would run the example but not check the output), but it doesn't matter much. Merging then!