Pandas Is there a count distinct in pandas?

Problem description

I can't seem to find if pandas has a count unique / distinct function. I know I can unique a series / dataframe and length that but it's quite slow. It would be neat if we had something that use hyperloglog since that's fast and memory efficient. Though it has a tradeoff of being probabilistic so that'd have to be configurable I feel.

Expected Output

I would expect something like

s = pd.Series([0,1,2])
s.count_distinct() #=> 3

Comment From: jorisvandenbossche

There is a nunique method, what I think returns what you are looking for. But, I think under the hood it just counts the uniques.

Comment From: hexgnu

awesome thanks @jorisvandenbossche